Intelligent control method for fermentation production of bacteriostatic agent based on pH and temperature coupling feedback

Abstract

This invention relates to the field of fermentation engineering control technology, specifically to an intelligent control method for the fermentation production of antimicrobial agents based on pH and temperature coupling feedback. The method involves establishing an offline prior knowledge base including a three-dimensional response surface model of pH, temperature, and product synthesis, as well as bidirectional coupling influence coefficients. Real-time data acquisition of current pH and temperature values ​​is performed via hardware synchronization, and the fermentation stage is automatically identified by combining auxiliary parameters such as dissolved oxygen and carbon dioxide release rates. The deviation of the current value from the optimal coupling range is calculated and superimposed with a coupling perturbation prediction value to generate a comprehensive control demand signal. Based on the effective state of the demand signal and the priority of the fermentation stage, either single-variable control with feedforward compensation or bivariate collaborative control based on stepped alternating output is executed. This invention achieves decoupled collaborative control of pH and temperature, effectively improving the fermentation potency of antimicrobial agents, shortening the fermentation cycle, reducing acid and alkali consumption and energy consumption, and enhancing batch-to-batch production consistency.

Description

Technical Field

[0001] This invention relates to the field of fermentation engineering control technology, and in particular to an intelligent control method for the fermentation production of antibacterial agents based on pH and temperature coupling feedback. Background Technology

[0002] Antimicrobial agents, such as nisin and natamycin, are a class of natural biological preservatives produced through microbial secondary metabolism, with broad application prospects in food, medicine, and other fields. Their fermentation production process is a complex nonlinear dynamic system, and the growth, metabolism, and antimicrobial agent synthesis of microorganisms are extremely sensitive to the physicochemical parameters of the culture environment. Among these, the pH and temperature of the fermentation broth are two of the most critical control parameters. Traditional antimicrobial agent fermentation control strategies typically treat pH and temperature as independent control loops, employing PID control or simple setpoint-hysteresis control methods. Specifically, a constant pH setpoint (e.g., 6.5) and a constant temperature setpoint (e.g., 30°C) are set separately. When the sensor detects that the actual value deviates from the set threshold, an acid / alkali feed pump or heating / cooling device is independently activated for adjustment. However, in actual fermentation processes, pH and temperature control exhibit a strong coupling effect: for example, the addition of acidic or alkaline solutions inevitably causes local temperature changes in the fermentation broth; and temperature changes not only affect microbial enzyme activity but also the solubility balance of acidic / alkaline substances such as CO2 and organic acids in the fermentation broth, thus significantly altering the pH value. Current technologies neglect this coupling relationship, leading to frequent interference between the two control loops and an "overshoot-correction" oscillation phenomenon. This not only increases acid and alkali consumption and energy consumption but, more importantly, fails to keep microorganisms consistently on the optimal product synthesis metabolic pathway, resulting in low antimicrobial agent fermentation potency, large batch-to-batch variability, and long fermentation cycles. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent control method for the fermentation production of antibacterial agents based on pH and temperature coupling feedback, which can sense and decouple the pH and temperature coupling effects in real time, thereby achieving synergistic and coordinated control between the two.

[0004] To achieve the above objectives, this invention provides a smart control method for the fermentation production of antibacterial agents based on pH and temperature coupling feedback, comprising the following steps: Through multi-factor cross-experiment, the specific synthesis rate of antibacterial agents under different combinations of pH and temperature values ​​was determined, a three-dimensional response surface model was constructed, the coupling combination corresponding to different fermentation stages was extracted from it, and the short-term influence coefficient of pH adjustment action on temperature and the short-term influence coefficient of temperature adjustment action on pH were determined and stored in the prior knowledge base. The current pH and temperature values ​​are collected synchronously and continuously, while auxiliary parameters are also collected. The current fermentation stage is automatically identified based on the collected data and auxiliary parameters. Based on the current fermentation stage, the corresponding optimal pH range and optimal temperature range are retrieved from the prior knowledge base. The calculated first deviation and the obtained first coupling perturbation estimate are superimposed to generate a comprehensive pH control demand signal. The calculated second deviation and the obtained second coupling perturbation estimate are superimposed to generate a comprehensive temperature control demand signal. Based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the priority rules preset for the current fermentation stage, a collaborative control instruction set is generated. The coordinated control instruction set is issued, and the current pH value and current temperature value after the coordinated control instruction set is executed are collected to form a closed-loop coupled feedback control loop.

[0005] The synchronous method continuously collects the current pH value and the current temperature value, including: At the beginning of each sampling cycle, a synchronization pulse signal is generated and sent to both the pH detection unit and the temperature detection unit. Upon receiving the synchronization pulse, both detection units latch the sensor signals and complete the analog-to-digital conversion. The converted data, along with the same timestamp, is then sent to the coupling feedback controller. The synchronization pulse signal is generated by the coupling feedback controller.

[0006] The fermentation stage includes the cell growth stage, the inhibitory substance synthesis stage, and the cell autolysis stage; the auxiliary parameters include dissolved oxygen concentration and carbon dioxide release rate; the coupled feedback controller has a built-in fuzzy inference engine for the fermentation stage, which takes the current pH value, current temperature value, first derivative of dissolved oxygen concentration, second derivative of carbon dioxide release rate, and fermentation time as input feature vectors, outputs the membership degree of each stage at the current moment, and uses a hysteresis comparator to suppress frequent jumps between stages, and finally outputs the current fermentation stage.

[0007] The coupling combination is determined based on the flat plateau region in the three-dimensional response surface model where the change in the rate of synthesis of antibacterial agent is less than a preset threshold; and different fermentation stages correspond to different optimal pH ranges and optimal temperature ranges. When switching stages, the coupling feedback controller adopts a linear transition strategy for the endpoint values ​​of the target range, so that the target range values ​​of the old stage change uniformly to the target range values ​​of the new stage.

[0008] The first coupling disturbance estimate is: the estimated pH change caused by a standard temperature adjustment action through the short-term influence coefficient. The second coupling disturbance estimate is: the estimated temperature change caused by a standard pH adjustment action through the short-term influence coefficient. The comprehensive pH control demand signal is equal to the first deviation minus the first coupling disturbance estimate, and the comprehensive temperature control demand signal is equal to the second deviation minus the second coupling disturbance estimate.

[0009] Before generating the coordinated control instruction set, the method further includes: When the normalized integrated pH control demand signal deviates from the zero demand benchmark by less than the pH dead zone threshold, the signal is marked as invalid; when the normalized integrated temperature control demand signal deviates from the zero demand benchmark by less than the temperature dead zone threshold, the signal is marked as invalid; a coordinated control instruction set is generated only when the signal is marked as valid.

[0010] Specifically, based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the preset priority rules of the current fermentation stage, a coordinated control instruction set is generated, including: When only the comprehensive pH control demand signal is valid while the comprehensive temperature control demand signal is invalid, the coupled feedback controller determines the feeding direction and acceleration rate of the acid-base feeding execution unit based on the comprehensive pH control demand signal. At the same time, it generates a short-term reverse compensation command based on the second coupled disturbance prediction value and sends it to the temperature regulation execution unit, so that the temperature regulation execution unit can simultaneously perform compensation actions to counteract the temperature disturbance while the acid-base is fed.

[0011] Specifically, based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the preset priority rules of the current fermentation stage, a coordinated control instruction set is generated, including: When only the comprehensive temperature control demand signal is valid while the comprehensive pH control demand signal is invalid, the coupled feedback controller determines the adjustment direction and adjustment power of the temperature control execution unit based on the comprehensive temperature control demand signal. At the same time, it generates a short-term reverse compensation command based on the first coupled disturbance prediction value and sends it to the acid-base flow execution unit, so that the acid-base flow execution unit can simultaneously perform compensation actions to counteract pH disturbances while adjusting the temperature.

[0012] Specifically, based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the preset priority rules of the current fermentation stage, a coordinated control instruction set is generated, including: When both the comprehensive pH control demand signal and the comprehensive temperature control demand signal are valid, the coupled feedback controller first retrieves the preset control priority weights according to the current fermentation stage, and then judges the degree of coupling conflict between the two control actions: if it is a weak conflict or mutually beneficial, it outputs a parallel full-intensity command, so that the two execution units execute the actions at their respective calculated intensities simultaneously; if it is a strong conflict, it executes a step-by-step alternating output strategy, dividing a control cycle into multiple sub-cycles. In each sub-cycle, the high-priority control action is executed first, and the low-priority control demand is attenuated and converted into a feedforward compensation amount and superimposed on the high-priority action command. After the sub-cycle ends, the roles are switched, and this alternation continues until both comprehensive demand signals drop below the dead zone.

[0013] The priority weights of the regulation are preset according to different fermentation stages: in the cell growth stage, the priority of temperature regulation is higher than that of pH regulation; in the inhibitory substance synthesis stage, the priority of pH regulation is higher than that of temperature regulation; and in the cell autolysis stage, the priority of temperature regulation is higher than that of pH regulation.

[0014] This invention discloses an intelligent control method for the fermentation production of antimicrobial agents based on pH and temperature coupling feedback. It establishes an offline prior knowledge base including a three-dimensional response surface model of pH, temperature, and product synthesis, as well as bidirectional coupling influence coefficients. Real-time hardware synchronization is used to collect current pH and temperature values, and auxiliary parameters such as dissolved oxygen and carbon dioxide release rates are used to automatically identify the fermentation stage. The deviation of the current value from the optimal coupling range is calculated and superimposed with the coupling perturbation prediction value to generate a comprehensive control demand signal. Then, based on the effective state of the demand signal and the priority of the fermentation stage, either single-variable control with feedforward compensation or bivariate collaborative control based on stepped alternating output is executed. This invention achieves decoupled collaborative control of pH and temperature, effectively improving the fermentation potency of antimicrobial agents, shortening the fermentation cycle, reducing acid and alkali consumption and energy consumption, and enhancing batch-to-batch production consistency. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below.

[0016] Figure 1 This is a schematic diagram of the steps of an intelligent control method for the fermentation production of antibacterial agents based on pH and temperature coupling feedback provided by the present invention.

[0017] Figure 2 This is a simplified flowchart illustrating an intelligent control method for the fermentation production of antibacterial agents based on pH and temperature coupling feedback, provided by this invention.

[0018] Figure 3This is a complete flowchart of an intelligent control method for the fermentation production of antibacterial agents based on pH and temperature coupling feedback, provided by the present invention. Detailed Implementation

[0019] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0020] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0021] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0022] Please see Figures 1-3 This invention provides a smart control method for the fermentation production of antibacterial agents based on pH and temperature coupling feedback, comprising the following steps: S1. Through multi-factor cross-experiment, the specific synthesis rate of antibacterial agent under different combinations of pH and temperature values ​​was determined, a three-dimensional response surface model was constructed, the coupling combination corresponding to different fermentation stages was extracted from it, and the short-term influence coefficient of pH adjustment action on temperature and the short-term influence coefficient of temperature adjustment action on pH were determined and stored in the prior knowledge base.

[0023] Specifically, before the fermentation process begins, systematic offline experiments are conducted to acquire all prior knowledge of the specific antimicrobial agent producing strain and its fermentation system. The two most significant environmental variables affecting antimicrobial agent fermentation are selected: pH and temperature of the fermentation broth. Based on the biological characteristics of the strain, the pH range (e.g., for nisin-producing strains, the pH range is set to 5.0–7.5, with intervals of 0.25) and the temperature range (e.g., 25℃–35℃, with intervals of 1.0℃) are determined. A central composite design or Box-Behnken design is used to form an experimental matrix containing at least 30 different (pH, temperature) combinations. Each combination corresponds to an independent batch fermentation experiment. For each (pH, temperature) combination in the experimental matrix, small-scale fermentation is conducted under strictly constant conditions. Throughout the fermentation cycle, aseptic sampling is performed at fixed time intervals (e.g., 1 hour), and the following three types of data are recorded: cell concentration: determined using optical density or dry weight method; substrate concentration (e.g., glucose): determined using a biosensor analyzer. Target antibacterial agent concentration: its biopotency or mass concentration was determined by agar diffusion method or high performance liquid chromatography.

[0024] Differentiate the time-series data for each batch of experiments and calculate: Cell growth rate (the amount of cells added per unit cell per unit time).

[0025] Substrate consumption rate.

[0026] The specific synthesis rate of the antibacterial agent (i.e., the amount of the target antibacterial agent synthesized per unit cell per unit time, denoted as q_p).

[0027] Among them, the rate of synthesis of antibacterial agents is the core indicator for evaluating the quality of the fermentation microenvironment, because it directly reflects the efficiency of the cells in producing the target product under the current pH and temperature coupling conditions.

[0028] For each experimental combination point, the pH value and temperature value are used as two input variables, and the maximum specific synthesis rate of the antimicrobial agent at that point (the peak value reached by this parameter during fermentation) is used as the response output value. A quadratic polynomial regression or radial basis function neural network is used to fit the data of all experimental points, resulting in a continuous and smooth three-dimensional response surface model. This model can quantitatively describe the theoretical maximum specific synthesis rate of the antimicrobial agent under any combination of pH and temperature values. This invention does not use a single optimal pH and optimal temperature setpoint, but rather identifies a flat plateau region in the response surface. Within this region, the variation in the specific synthesis rate of the antimicrobial agent is less than a preset threshold (e.g., less than 5%). This region is defined as the optimal pH-temperature coupling interval, rather than a single point. The reason for using an interval instead of a point is that, in actual fermentation, due to perturbations and measurement noise, it is almost impossible to simultaneously and precisely control both parameters at a single point, while accepting an acceptable interval can greatly improve the robustness and feasibility of regulation.

[0029] The aforementioned three-dimensional response surface model reflects the influence of the "instantaneous" environment on the product synthesis rate. However, during the complete fermentation cycle, the cells undergo different stages such as growth, synthesis, and death, and their metabolic requirements change accordingly. Therefore, this invention further performs the following operations: For each fermentation stage (cell growth stage, bacteriostatic agent synthesis stage, and cell autolysis stage), the above multi-factor crossover experiment was repeated. Specifically, for the growth stage, the effects of pH and temperature on the specific growth rate of the cells were investigated to find the coupling range that maximizes the growth rate; for the bacteriostatic agent synthesis stage, the effects on the specific synthesis rate of the bacteriostatic agent were investigated to find the coupling range that maximizes product synthesis; for the autolysis stage, the effects on the activity of cell lysin and the product release rate were investigated to find the coupling range that delays autolysis and maintains product stability.

[0030] Thus, an optimal pH-temperature coupled trajectory is obtained that varies with fermentation time. This trajectory is not two independent pH setpoint curves and temperature setpoint curves, but a two-dimensional interval sequence: at each fermentation time point or time window, there is an optimal pH range and an optimal temperature range, and there is an inherent correspondence between these two ranges (for example, in the synthesis stage, a higher pH corresponds to a lower suitable temperature, and vice versa, forming a "ridge line").

[0031] To decouple the mutual interference between pH and temperature during the regulation process, this invention specifically designs a perturbation experiment to quantitatively determine the short-term influence coefficients in both directions.

[0032] Determining the short-term effect coefficient of pH adjustment on temperature: In a fermenter maintaining a stable temperature (e.g., 30°C), a certain amount (e.g., 10 mL) of acid or alkali solution is added at a standard rate. Simultaneously, a high-precision temperature sensor records the temperature change within the fermenter over the following 30 seconds. Due to the release or absorption of heat during acid-base neutralization or dissolution, a brief temperature rise or fall occurs. The experiment is repeated multiple times, changing conditions such as the buffer capacity of the fermentation broth and the cell concentration, to fit the temperature change caused by a unit acid / base addition. This coefficient is denoted as K_acid-base-temperature, and since it is a function of the current temperature, current pH value, and cell concentration, it is stored in a lookup table.

[0033] Determining the short-term effect coefficient of temperature regulation on pH: In a fermentation broth with a stable pH in a fermenter, the heating or cooling device was started at standard power to raise or lower the temperature by 1°C. Simultaneously, the pH change was recorded over the next 30 seconds using a high-precision pH meter. Because temperature changes affect CO2 solubility, the dissociation constant of weak acids / weak bases, and the rate of organic acid production by cell metabolism, a brief pH shift occurs. The experiment was repeated to fit the pH change caused by a unit temperature change. This coefficient is denoted as K_temperature versus pH and is also stored in a multidimensional table format.

[0034] Ultimately, the prior knowledge base is a structured database stored in the non-volatile memory of the coupled feedback controller, containing at least the following data types, as shown in Table 1: Table 1 All the above data were obtained through offline experiments and stored in the controller before fermentation began. This prior knowledge base can be reused for the same production process of the same strain, without needing to be rebuilt for each batch; it only needs to be fine-tuned and calibrated periodically based on production data.

[0035] S2. Simultaneously collect the current pH value and current temperature value, and collect auxiliary parameters. Based on the collected data and auxiliary parameters, automatically identify the current fermentation stage.

[0036] Specifically, pH and temperature values ​​change dynamically over time during fermentation, and there is a strong coupling between them. If the pH and temperature detection units are not synchronized—for example, if the temperature data is recorded a few seconds later than the pH data—the coupled feedback controller will incorrectly combine the environmental states from the two different time points when calculating the current overall deviation, creating a "false current state." This temporal misalignment can lead to the following consequences: The coupling compensation factor calculation is distorted because the coupling effect coefficient (such as the short-term effect of temperature regulation on pH) is based on the pH and temperature values ​​at the same time. Time misalignment will cause the table lookup index to be incorrect.

[0037] Lagging or premature regulatory decisions: Controllers may make incorrect judgments about the changed environment, triggering unnecessary adjustments or missing the best opportunity for adjustment.

[0038] Therefore, this invention requires that the pH detection unit and the temperature detection unit must be strictly synchronized under the same clock reference, that is, the two sensors trigger data reading at the same time and stamp the reading results with the same timestamp.

[0039] The coupled feedback controller integrates a high-precision real-time clock module and a hardware-level synchronization trigger circuit. The specific implementation process is as follows: Clock synchronization reference: All acquisition units (pH detection unit, temperature detection unit, and auxiliary parameter detection unit described below) use the real-time clock module within the coupled feedback controller as their sole time source. Before fermentation begins, the controller broadcasts a synchronization calibration command to each detection unit via an industrial Ethernet or dedicated synchronization signal line to eliminate minor deviations in the crystal oscillators within each unit.

[0040] Periodic synchronous triggering: The coupled feedback controller is set to a fixed sampling period, such as 10 seconds. At the beginning of each sampling period, the controller's internal hardware timer generates a synchronization pulse signal. This signal is simultaneously sent to the digital interfaces of the pH detection unit and the temperature detection unit via a signal distributor.

[0041] Data latching and transmission: Upon receiving the synchronization pulse, the pH and temperature detection units immediately latch the current analog signals from the sensors, which are then converted by their respective analog-to-digital converters. The converted digital values ​​are encapsulated in a data frame containing: the detection unit identification code, the sampled value, and the period number of this sampling. Since all units are triggered by the same pulse, their data naturally correspond to the same moment.

[0042] Timestamp marking: When the coupled feedback controller sends out a synchronization pulse, it records the current real-time clock value as the absolute timestamp for that sampling period. When the controller receives data frames returned by each detection unit, it appends this absolute timestamp to the frame to form a complete synchronization sampling record.

[0043] Through the aforementioned hardware and software synchronization mechanism, this invention ensures that at any given time, the pH and temperature values ​​processed by the controller are a true reflection of the same microenvironmental state within the fermenter.

[0044] The fermentation stage is a macroscopic division of the physiological state of the microorganisms throughout the fermentation process. For the fermentation production of antibacterial agents, this invention divides the complete fermentation cycle into the following three standard stages, each with clear biological markers and process characteristics, as shown in Table 2: Table 2 The division of these three stages is not based on fixed time points, because differences in strain activity, raw material batches, and inoculum size between different batches can cause the start and end times of each stage to drift. Therefore, this invention employs a multi-parameter fusion-based stage identification method, rather than relying solely on fermentation timing.

[0045] Auxiliary parameters refer to process variables that reflect the physiological state of bacteria, in addition to pH and temperature. This invention collects at least the following three auxiliary parameters: Dissolved oxygen concentration: measured in real time by dissolved oxygen electrodes installed inside the fermenter.

[0046] Carbon dioxide emission rate: detected online by an exhaust gas analyzer.

[0047] Fermentation time: Recorded by the controller's internal timer.

[0048] The recognition process is as follows: Step 1: Data Synchronization Acquisition and Preprocessing: The auxiliary parameter detection unit mentioned above also participates in the synchronous acquisition process in Step 2, that is, it is triggered under the same synchronization pulse as the pH and temperature detection units to obtain dissolved oxygen concentration values ​​and carbon dioxide release rate values ​​with the same timestamp. The coupled feedback controller performs moving average filtering on the raw data to remove high-frequency noise.

[0049] Step 2: Extracting the Feature Vector for Stage Identification: At each sampling moment, the controller assembles a set of data at the current timestamp into a feature vector, including: the current pH value, the current temperature value, the first derivative of the dissolved oxygen concentration (reflecting its changing trend), the second derivative of the carbon dioxide release rate (reflecting its accelerating changing trend), and the current fermentation time. The derivatives are calculated using the time-difference method, but it is important to ensure matching with the synchronous sampling period.

[0050] Step 3: Stage discrimination based on fuzzy rules: This invention pre-configures a fermentation-stage fuzzy inferencer within the coupled feedback controller. Its input is the aforementioned feature vector, and its output is the membership degree (between 0 and 1) of each stage at the current moment. Examples of inference rules are as follows: The rule for determining the stage of bacterial growth is as follows: if the first derivative of dissolved oxygen concentration is negative and has a large absolute value (indicating that dissolved oxygen is being consumed rapidly), and the second derivative of carbon dioxide release rate is positive (indicating that the release rate is accelerating), and the rate of antibacterial agent synthesis (estimated by an offline soft measurement model, or not used if it is not yet known) has not increased significantly, then the bacterial growth stage is considered to have a high degree of membership.

[0051] The criteria for identifying the inhibitory substance synthesis stage are as follows: if the first derivative of dissolved oxygen concentration changes from negative to positive and the absolute value is very small (i.e., dissolved oxygen "rebound inflection point"), and the second derivative of carbon dioxide release rate changes from positive to negative (i.e., the release rate starts to slowly decline after reaching its peak), and the current fermentation time has exceeded the minimum growth stage duration (e.g., 6 hours), then it is determined to be the inhibitory substance synthesis stage.

[0052] The criteria for determining the autolysis stage are as follows: if the first derivative of the dissolved oxygen concentration is positive and continues to increase (indicating a sharp decrease in oxygen consumption), and the first derivative of the carbon dioxide release rate is negative and its absolute value increases (indicating a significant decline in metabolic activity), and the fermentation time exceeds the historical average duration of the synthesis stage, then it is determined to be in the autolysis stage.

[0053] Step 4: Output the final recognition results: Of the three membership degrees output by the fuzzy inference engine, the stage corresponding to the highest value is taken as the identification result of the current fermentation stage. To prevent frequent jumps between stages (e.g., repeated switching at boundaries), the controller introduces a hysteresis comparator: a stage switch is only allowed when the membership degree of a new stage exceeds a certain threshold (e.g., 0.3) of the current stage's membership degree for more than three consecutive sampling periods. After the switch is completed, the timestamp of the stage switch is recorded and used for the duration statistics of subsequent stages.

[0054] Step 5: Smooth Transition of Coupled Targets During Stage Switching: When a change in the fermentation stage is detected (e.g., switching from the cell growth stage to the inhibitory substance synthesis stage), the corresponding optimal pH-temperature coupled target range will also change abruptly. To avoid drastic actuator movements due to sudden changes in the target, this invention adopts a linear transition strategy: within a 5-minute time window before and after the stage switching point, the endpoint value of the target range output by the coupled feedback controller changes uniformly from the value of the old stage to the value of the new stage, rather than undergoing a step change.

[0055] S3. Based on the current fermentation stage, retrieve the corresponding optimal pH range and optimal temperature range from the prior knowledge base. Superimpose the calculated first deviation and the obtained first coupling perturbation estimate to generate a comprehensive pH control demand signal. Superimpose the calculated second deviation and the obtained second coupling perturbation estimate to generate a comprehensive temperature control demand signal.

[0056] Specifically, based on the identified current fermentation stage (cell growth stage, bacteriostatic agent synthesis stage, or cell autolysis stage), the optimal coupling trajectory table in the prior knowledge base is accessed. This table stores the corresponding optimal pH and temperature ranges for each fermentation stage. For reference, for typical nisin-producing strains, the target ranges for these three stages can be set as shown in Table 3 (for reference only; actual values ​​vary depending on the strain): Table 3 All of the above intervals are closed intervals. It is important to emphasize that these two intervals are not independent, but rather have an inherent correspondence: for example, during the bacteriostatic synthesis stage, when the pH value is at the lower limit of the interval (6.0), the optimal temperature interval will shift to near its upper limit (30.5℃), and vice versa. This correspondence is implicitly represented as interval ridges in the response surface model of the prior knowledge base, but in this step, the interval boundaries are directly used as control targets to simplify the online computation.

[0057] The current pH value sent by the current pH detection unit is compared with the optimal pH range to calculate the first deviation. The calculation method is as follows: If the current pH value is within the optimal pH range (including the endpoints), then the first deviation is zero.

[0058] If the current pH value is lower than the lower limit of the optimal pH range, the first deviation is "the lower limit value minus the current pH value", which is recorded as a positive deviation (indicating that the pH value needs to be adjusted upward).

[0059] If the current pH value is higher than the upper limit of the optimal pH range, the first deviation is "the current pH value minus the upper limit value", which is recorded as a positive deviation (indicating that the pH value needs to be lowered).

[0060] Similarly, the current temperature value is compared with the optimal temperature range, and the second deviation is calculated according to the same rules. For example, if the current temperature is 31.5℃, and the optimal temperature range for the inhibitor synthesis stage is 28.5℃~30.5℃, then the current temperature is 1.0℃ higher than the upper limit, and the second deviation is 1.0℃.

[0061] The coupling compensation factor is a quantitative description of the physicochemical coupling effect of pH regulation affecting temperature and temperature regulation affecting pH. This invention obtains these factors using the bidirectional coupling influence coefficients measured offline in step one. Specifically, two lookup tables are stored in a priori knowledge base: pH→Temperature Influence Coefficient Table: This table records the short-term temperature change of the fermentation broth caused by a unit acid / base feed rate (e.g., 1 mL of 1 mol / L sodium hydroxide solution per liter of fermentation broth) under given current pH, temperature, and cell concentration (estimated by an offline model or obtained through online soft sensing). This change can be positive or negative: alkali feed is usually exothermic, causing a temperature increase, while acid feed may be endothermic, causing a temperature decrease. This coefficient is denoted as K_acid-base-temperature.

[0062] Temperature-pH Effect Coefficient Table: This table records the short-term pH change of the fermentation broth caused by a unit temperature change (e.g., a 1°C change) under given current pH, temperature, and cell concentration conditions. This change is typically negative: increased temperature decreases carbon dioxide solubility and alters the dissociation constant of weak acids, often leading to a decrease in pH; conversely, decreased temperature increases pH. This coefficient is denoted as K_temperature versus pH.

[0063] Based on the current pH and temperature values, the two coefficient tables mentioned above are consulted to obtain the two coefficient values. Then, it is necessary to estimate: how much disturbance to the temperature would occur if a standard pH adjustment were immediately performed, and how much disturbance to the pH would occur if a standard temperature adjustment were immediately performed. Specifically, a standard adjustment equivalent is preset. For example: The standard pH adjustment equivalent is defined as the volume of acid or alkali solution required to change the pH of 1 cubic meter of fermentation broth by 0.1 units (this volume can be pre-calculated using the fermentation broth buffer capacity).

[0064] The standard temperature regulation equivalent is defined as the amount of heat required to change the temperature of 1 cubic meter of fermentation broth by 0.5 degrees Celsius (corresponding to the standard operating time of the heating or cooling device, such as 30 seconds).

[0065] then: First Coupled Perturbation Estimation (Expected Effect of Temperature Regulation on pH) = Standard Temperature Regulation Equivalent × K_Temperature to pH. This estimate indicates how much additional pH value will change if a standard temperature regulation is initiated.

[0066] The second coupled perturbation estimate (expected effect of pH adjustment on temperature) = Standard pH adjustment equivalent × K_acid-base-temperature. This estimate indicates how much additional temperature change will occur if a standard pH adjustment is initiated.

[0067] These two estimates have positive and negative signs. For example, K_temperature is usually negative for pH (heating causes pH to decrease), so the first coupled perturbation estimate is usually negative, indicating that if the temperature increases, the pH will decrease. The second coupled perturbation estimate depends on whether acid or base is added: if the pH needs to be increased (adding base), then K_acid-base is positive for temperature (exothermic), and the perturbation estimate is positive, indicating that the temperature will increase.

[0068] The basic deviation is superimposed with the corresponding coupled disturbance estimate to generate a comprehensive control demand signal. The direction of superposition is determined by the control logic: the coupled disturbance estimate is introduced to offset the side effects of the impending control action on another parameter, so it should be in the same or opposite direction as the basic deviation, depending on whether the disturbance is helpful or harmful to achieving the target.

[0069] The specific rules are as follows: Overall pH regulation demand signal = First deviation + (First coupling perturbation estimate × Coupling direction coefficient) The coupling direction coefficient is either +1 or -1. The judgment logic is as follows: if the sign of the first coupling disturbance estimate is the same as the first deviation, it means that the direction of pH change caused by temperature regulation is consistent with the correction direction of the current pH deviation (i.e., the disturbance helps pH return to the target range). In this case, the disturbance is "beneficial," and no pH regulation is needed for compensation; instead, it reduces the need for pH regulation. Therefore, the first coupling disturbance estimate is multiplied by -1 before being added, effectively subtracting the beneficial disturbance from the basic deviation. Conversely, if the sign of the first coupling disturbance estimate is opposite to the first deviation (the disturbance causes pH to deviate further from the target), the disturbance is "harmful" and requires additional compensation. Therefore, it is multiplied by +1 and directly added. To simplify the description without loss of generality, this invention adopts the most robust feedforward compensation strategy: regardless of whether the disturbance is beneficial or harmful, it is added as a feedforward term to the comprehensive demand signal, with the direction of canceling out the disturbance. That is: Comprehensive pH regulation demand signal = First deviation amount - First coupling disturbance estimate.

[0070] The physical meaning is as follows: If temperature regulation is about to cause a pH increase, and the current pH is already too high (needing to decrease), then the overall demand should be greater than the baseline deviation to generate a stronger downward adjustment instruction to offset the upward disturbance. If temperature regulation is about to cause a pH increase, and the current pH is too low (needing to increase), then the overall demand should be less than the baseline deviation, or even become negative (indicating that pH regulation is not needed for the time being, and reverse regulation may even be allowed). In actual implementation, the overall demand signal is limited to between 0 and 1, with negative values ​​set to 0.

[0071] Comprehensive temperature control demand signal = Second deviation - Second coupling disturbance estimate.

[0072] Similarly, the second coupling disturbance estimate is subtracted from the second deviation to pre-cancele the temperature disturbance caused by pH adjustment.

[0073] To avoid frequent actuator movements due to measurement noise and minor fluctuations, this invention sets a start-up dead-zone threshold for each comprehensive control demand signal. The reference setting is as follows: pH control dead zone threshold: corresponding to an equivalent deviation of ±0.05 pH units.

[0074] Temperature control dead zone threshold: corresponding to an equivalent deviation of ±0.2 degrees Celsius.

[0075] When the normalized value of the integrated control demand signal deviates from 0.5 by less than the dead zone threshold (e.g., between 0.48 and 0.52), it is considered that there is no effective control demand, and the signal is marked as invalid. Only when it exceeds the dead zone threshold is the signal considered as effective demand, and it is used for joint decision-making in step four.

[0076] S4. Based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and in conjunction with the priority rules preset for the current fermentation stage, generate a collaborative control instruction set.

[0077] Specifically, the collaborative decision-maker inside the coupled feedback controller receives two signals from the real-time decoupling decision-making module in step three: the comprehensive pH control demand value (denoted as R_pH) and the comprehensive temperature control demand value (denoted as R_T). Both signals are normalized to the range of 0 to 1, where 0.5 represents no net control demand, greater than 0.5 indicates that pH needs to be increased (i.e., adding alkali) or temperature needs to be raised, and less than 0.5 indicates that pH needs to be decreased (i.e., adding acid) or temperature needs to be lowered. Simultaneously, each signal is accompanied by a validity flag, which is determined by the dead zone threshold in step three: when |R_pH - 0.5| is less than the dead zone threshold (e.g., 0.02, corresponding to an equivalent pH deviation of 0.05), the validity flag is "invalid"; otherwise, it is "valid". Similarly, the threshold for the temperature validity flag corresponds to an equivalent temperature deviation of 0.2℃ (approximately 0.02 after normalization).

[0078] The decision-maker first determines the state of two valid flags, and then proceeds to the corresponding decision branch based on different combinations. In each branch, the decision-maker calculates specific execution instruction parameters, including: the direction of acid or alkali addition (acid or alkali), flow rate (mL / min), and expected duration for the acid-base addition execution unit; and the direction of temperature regulation (heating or cooling), regulation power (watts or percentage of rated power), and expected duration for the temperature regulation execution unit. All instructions are encapsulated into a coordinated control instruction set, which is sent to the execution unit via the controller output module.

[0079] Each case will be explained in detail below.

[0080] Scenario 1: Both comprehensive control demand signals are ineffective Judgment criteria: The valid indicator of the comprehensive pH regulation demand signal is "invalid" and the valid indicator of the comprehensive temperature regulation demand signal is "invalid".

[0081] Decision logic: The current pH and temperature values ​​in the fermentation environment are both within their optimal target ranges, and the net demand after considering feedforward compensation for coupled perturbations is close to zero; therefore, no active adjustment is required. The decision-maker outputs a maintenance instruction set, specifically including: Send a "standby" command to the acid / base feed unit to stop all feed pumps and maintain the current valve status. Send a "hold" command to the temperature control unit to stop the active input of heating or cooling media and maintain the current temperature solely through the tank insulation layer.

[0082] Scenario 2: Only effective based on the overall pH regulation demand signal Judgment criteria: The overall pH regulation demand signal is valid, while the overall temperature regulation demand signal is invalid.

[0083] Decision logic: At this point, only the pH value deviates from the optimal range (or its net deviation after coupled feedforward compensation exceeds the dead zone), while the temperature value is basically within acceptable limits. The decision-maker will perform univariate pH adjustment, but it must also consider that the pH adjustment action will cause a disturbance to the temperature (i.e., the second coupled disturbance prediction value). Therefore, feedforward compensation is needed for this disturbance to prevent the temperature from subsequently deviating from the range. The specific decision steps are as follows: Step 1: Determine the direction of pH adjustment and the baseline rate Based on the magnitude of R_pH relative to 0.5: If R_pH>0.5+ dead zone, then the pH needs to be adjusted upward, i.e., an alkaline solution needs to be added to the fermenter.

[0084] If R_pH < 0.5 - dead zone, then the pH needs to be lowered, i.e., acid solution needs to be added to the fermenter.

[0085] The baseline flow acceleration rate is determined by the degree to which R_pH deviates from 0.5; the greater the deviation, the faster the rate. For example, a linear proportional relationship is used: when R_pH = 0.7, the flow acceleration rate is set to 50% of the maximum allowable rate (e.g., 100 mL / min); when R_pH is close to the dead zone boundary, the flow acceleration rate is set to 10% of the minimum adjustable rate. The specific rate mapping curves are pre-stored in a table format in the prior knowledge base and can be adjusted according to different strains and fermentation scales.

[0086] Step 2: Calculate the temperature feedforward compensation. The decision-maker retrieves the estimated second coupled disturbance caused by the current pH adjustment action from the calculation results in step three (i.e., the temperature change caused by a unit standard pH adjustment equivalent, multiplied by the current actual required adjustment equivalent). Since this disturbance estimate was calculated based on the standard equivalent in step three, this step needs to convert it into a feedforward compensation command for the temperature regulation execution unit. Specifically, while issuing the pH adjustment command, the decision-maker sends a short-term reverse compensation command to the temperature regulation execution unit. For example, if increasing the pH (adding alkali) will cause the fermentation broth temperature to rise by 0.2°C, the decision-maker instructs the temperature regulation execution unit to perform an equivalent cooling action within the same time window (e.g., running at 15% of rated power for 10 seconds) to offset the temperature rise. The duration of this compensation command is synchronized with the expected duration of the pH adjustment, and the compensation strength is equal to the estimated coupled disturbance divided by the standard regulation capacity of the temperature regulation unit.

[0087] Step 3: Generate the coordinated control instruction set This instruction set contains two parallel instructions: Acid-base feeding execution unit: Performs feeding in the calculated direction and rate for a duration of a preset adjustment period (e.g., 30 seconds), and re-triggers step two after the expiration.

[0088] Temperature regulation execution unit: synchronously executes feedforward compensation commands (heating or cooling, intensity and time matched with disturbance), and automatically returns to standby state after compensation is completed.

[0089] Scenario 3: Only the comprehensive temperature control demand signal is valid Judgment criteria: The overall temperature control demand signal is valid, while the overall pH control demand signal is invalid.

[0090] Decision logic: At this point, only the temperature value deviates from the optimal range, while the pH value is basically within acceptable limits. The decision-maker executes univariate temperature control and simultaneously performs feedforward compensation for pH perturbations. The specific steps are symmetrical to those in case two: Step 1: Determine the direction of temperature adjustment and the base power Based on the size of R_T relative to 0.5: If R_T > 0.5 + dead zone, heating is required (start heating).

[0091] If R_T < 0.5 - dead zone, cooling is required (start cooling).

[0092] The base power is linearly determined by the degree to which R_T deviates from 0.5; the greater the deviation, the higher the power.

[0093] Step 2: Calculate the pH feedforward compensation. The decision-maker retrieves the first coupled perturbation estimate calculated in step three (i.e., the expected impact of temperature regulation on pH). For example, increasing temperature typically leads to a decrease in pH. Therefore, if a temperature increase is required, and the current pH is within the target range, the temperature increase will lower the pH and may exceed the lower limit of the range. The decision-maker sends a short-term reverse compensation instruction to the acid-base feeding execution unit: if the temperature increase causes a decrease in pH, a micro-addition of alkali is performed simultaneously to maintain pH stability. The amount of alkali to be compensated is determined by the magnitude of the first coupled perturbation estimate and the fermentation broth buffer capacity.

[0094] Step 3: Generate the coordinated control instruction set Temperature regulation actuator: Performs heating or cooling with calculated direction and power, continuously for one regulation cycle.

[0095] Acid-base feedforward execution unit: synchronously executes pH feedforward compensation commands (adding acid or alkali, with intensity and time matching), and automatically goes into standby mode after compensation is completed.

[0096] Scenario 4: Both comprehensive regulatory demand signals are effective Judgment criteria: The valid criteria for both pH and temperature are "valid", that is, both deviate from the optimal range or their coupled net demand exceeds the dead zone.

[0097] Decision-making logic: This is the most complex but also the most critical situation. The decision-maker needs to handle two biases simultaneously and must avoid conflicting or overly competitive adjustment actions. This invention employs a three-layer strategy: stage priority rules + vectorized collaboration + tiered alternating output.

[0098] First level: Determining phase priorities The decision-maker retrieves the regulatory priority weights for the current fermentation stage from the prior knowledge base. Taking typical antimicrobial agent fermentation as an example: During the bacterial growth stage: temperature is more sensitive to bacterial proliferation than pH, therefore temperature regulation has a higher priority.

[0099] During the inhibitory agent synthesis stage: pH has a significantly higher weight in influencing the product synthesis rate than temperature, therefore pH regulation has a higher priority (usually set to 0.7, and temperature priority to 0.3).

[0100] During the autolysis stage: both need to remain stable, but rapid temperature fluctuations will accelerate autolysis, so temperature has a slightly higher priority.

[0101] Priority rules determine which action takes precedence when two regulatory actions conflict.

[0102] Second layer: Calculate the comprehensive control vector The decision-maker treats R_pH and R_T as vectors in a two-dimensional regulatory space. First, it determines whether the two regulatory directions are consistent or conflicting. Consistency of Direction Judgment: pH increase (adding alkali) is usually accompanied by temperature increase, while temperature increase (heating) is usually accompanied by pH decrease. Therefore, if both pH and temperature need to be increased simultaneously, the temperature rise caused by pH increase is consistent with the demand for temperature increase (beneficial), while the pH decrease caused by temperature increase is opposite to the demand for pH increase (conflicting). The decision-maker determines the overall degree of conflict by calculating the net coupling effect of the two adjustment actions. Specifically, the decision-maker substitutes the two coupling perturbation estimates calculated in step three: if the second coupling perturbation estimate (the effect of pH on temperature) is in the same direction as the temperature demand, and the first coupling perturbation estimate (the effect of temperature on pH) is opposite to the pH demand, it is judged as "strong conflict"; if both are in the same direction or both are opposite, it is judged as "weak conflict" or "beneficial".

[0103] Third layer: Generating cooperative instructions Different output strategies are adopted based on the degree and priority of conflict: In cases of weak conflict or mutual benefit: the two regulating actions can be executed simultaneously and complement each other. The decision-maker outputs parallel full-intensity instructions: the acid-base feeding unit feeds at the rate calculated by R_ph; the temperature regulating unit regulates the power calculated by R_T. In this case, no additional compensation is needed because the coupling effect of the two actions partially cancels out or adds up to within an acceptable range.

[0104] In cases of strong conflict, simultaneously performing two actions can lead to over-adjustment or oscillation. In this situation, the decision-maker employs a tiered alternating output strategy, which is one of the key innovations of this invention. Specifically: Divide an adjustment cycle (e.g., 60 seconds) into two sub-cycles, each lasting 30 seconds.

[0105] In the first sub-cycle, high-priority adjustment actions are executed first to output at full intensity; at the same time, low-priority actions are not output, but their demand intensity is attenuated and converted into a feedforward compensation amount, which is added to the instruction of the high-priority action (for example, if pH priority is high, the alkali addition rate is appropriately increased or decreased according to the temperature demand when adding alkali to take into account the temperature trend).

[0106] During the second sub-cycle, roles are switched: low-priority actions are performed, and feedforward compensation for high-priority actions is received.

[0107] Repeat this process alternately until both combined demand signals drop below the dead zone.

[0108] For example, during the bacteriostatic agent synthesis stage (where pH has high priority), alkali needs to be added (R_pH=0.7) and cooling is required (R_T=0.3, meaning the current temperature is too high). The strong conflict manifests as follows: adding alkali releases heat, raising the temperature, which conflicts with the cooling requirement. The decision-maker executes: for the first 30 seconds, alkali is added at full rate, while the rate is appropriately reduced based on the cooling requirement (because adding alkali hinders cooling), and cooling is initiated for slight feedforward compensation; for the next 30 seconds, alkali addition is stopped, and cooling is performed at full power, while simultaneously determining whether a small amount of alkali compensation is needed based on the pH requirement (which has been partially met). This process is repeated to achieve a gradual convergence of the two objectives.

[0109] Generate a coordinated control instruction set: Regardless of the sub-case described above, the final instruction set explicitly contains the action sequence (start time, duration, intensity, and direction) of the two execution units on the time axis. For stepped alternating output, the instruction set may contain multiple time segments.

[0110] S5. Issue the coordinated control instruction set, and simultaneously collect the current pH value and current temperature value after executing the coordinated control instruction set to form a closed-loop coupled feedback control loop.

[0111] Specifically, the coordinated control instruction set is sent to the acid-base feeding execution unit and the temperature regulation execution unit, which then receive and execute the coordinated control instruction set. Subsequently, the pH detection unit and the temperature detection unit collect new current values ​​again and return to step two, forming a continuous, closed-loop coupled feedback control cycle until fermentation ends.

[0112] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0113] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. An intelligent control method for fermentation production of bacteriostatic agent based on pH and temperature coupling feedback, characterized in that, Includes the following steps: Through multi-factor cross-experiment, the specific synthesis rate of antibacterial agents under different combinations of pH and temperature values ​​was determined, a three-dimensional response surface model was constructed, the coupling combination corresponding to different fermentation stages was extracted from it, and the short-term influence coefficient of pH adjustment action on temperature and the short-term influence coefficient of temperature adjustment action on pH were determined and stored in the prior knowledge base. The current pH and temperature values ​​are collected synchronously and continuously, while auxiliary parameters are also collected. The current fermentation stage is automatically identified based on the collected data and auxiliary parameters. Based on the current fermentation stage, the corresponding optimal pH range and optimal temperature range are retrieved from the prior knowledge base. The calculated first deviation and the obtained first coupling perturbation estimate are superimposed to generate a comprehensive pH control demand signal. The calculated second deviation and the obtained second coupling perturbation estimate are superimposed to generate a comprehensive temperature control demand signal. Based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the priority rules preset for the current fermentation stage, a collaborative control instruction set is generated. The coordinated control instruction set is issued, and the current pH value and current temperature value after the coordinated control instruction set is executed are collected to form a closed-loop coupled feedback control loop.

2. The intelligent control method for bacteriostat fermentation production based on pH and temperature coupling feedback according to claim 1, wherein, The system continuously collects the current pH and temperature values ​​in a synchronous manner, including: At the beginning of each sampling cycle, a synchronization pulse signal is generated and sent to both the pH detection unit and the temperature detection unit. Upon receiving the synchronization pulse, both detection units latch the sensor signals and complete the analog-to-digital conversion. The converted data, along with the same timestamp, is then sent to the coupling feedback controller. The synchronization pulse signal is generated by the coupling feedback controller.

3. The intelligent control method for antibacterial agent fermentation production based on pH and temperature coupling feedback as described in claim 1, characterized in that, The fermentation stage includes the cell growth stage, the inhibitory substance synthesis stage, and the cell autolysis stage; the auxiliary parameters include dissolved oxygen concentration and carbon dioxide release rate; the coupled feedback controller has a built-in fermentation stage fuzzy inference engine, which takes the current pH value, current temperature value, first derivative of dissolved oxygen concentration, second derivative of carbon dioxide release rate, and fermentation time as input feature vectors, outputs the membership degree of each stage at the current moment, and uses a hysteresis comparator to suppress frequent jumps between stages, and finally outputs the current fermentation stage.

4. The intelligent control method for antibacterial agent fermentation production based on pH and temperature coupling feedback as described in claim 1, characterized in that, The coupling combination is determined based on the flat plateau region in the three-dimensional response surface model where the change in the rate of synthesis of antibacterial agents is less than a preset threshold; and different fermentation stages correspond to different optimal pH ranges and optimal temperature ranges. When switching stages, the coupling feedback controller adopts a linear transition strategy for the endpoint values ​​of the target range, so that the target range values ​​of the old stage change uniformly to the target range values ​​of the new stage.

5. The intelligent control method for antibacterial agent fermentation production based on pH and temperature coupling feedback as described in claim 1, characterized in that, The first coupling disturbance estimate is: the estimated pH change caused by a standard temperature adjustment action through the short-term influence coefficient. The second coupling disturbance estimate is: the estimated temperature change caused by a standard pH adjustment action through the short-term influence coefficient. The comprehensive pH control demand signal is equal to the first deviation minus the first coupling disturbance estimate, and the comprehensive temperature control demand signal is equal to the second deviation minus the second coupling disturbance estimate.

6. The intelligent control method of bacteriostat fermentation production based on pH and temperature coupling feedback according to claim 1, wherein, Before generating the coordinated control instruction set, the method further includes: When the normalized integrated pH control demand signal deviates from the zero demand benchmark by less than the pH dead zone threshold, the signal is marked as invalid; when the normalized integrated temperature control demand signal deviates from the zero demand benchmark by less than the temperature dead zone threshold, the signal is marked as invalid; a coordinated control instruction set is generated only when the signal is marked as valid.

7. The intelligent control method for antibacterial agent fermentation production based on pH and temperature coupling feedback as described in claim 1, characterized in that, Based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the preset priority rules of the current fermentation stage, a coordinated control instruction set is generated, including: When only the comprehensive pH control demand signal is valid while the comprehensive temperature control demand signal is invalid, the coupled feedback controller determines the feeding direction and acceleration rate of the acid-base feeding execution unit based on the comprehensive pH control demand signal. At the same time, it generates a short-term reverse compensation command based on the second coupled disturbance prediction value and sends it to the temperature regulation execution unit, so that the temperature regulation execution unit can simultaneously perform compensation actions to counteract the temperature disturbance while the acid-base is fed.

8. The intelligent control method for antibacterial agent fermentation production based on pH and temperature coupling feedback as described in claim 1, characterized in that, Based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the preset priority rules of the current fermentation stage, a coordinated control instruction set is generated, including: When only the comprehensive temperature control demand signal is valid while the comprehensive pH control demand signal is invalid, the coupled feedback controller determines the adjustment direction and adjustment power of the temperature control execution unit based on the comprehensive temperature control demand signal. At the same time, it generates a short-term reverse compensation command based on the first coupled disturbance prediction value and sends it to the acid-base flow execution unit, so that the acid-base flow execution unit can simultaneously perform compensation actions to counteract pH disturbances while adjusting the temperature.

9. The intelligent control method for antibacterial agent fermentation production based on pH and temperature coupling feedback as described in claim 1, characterized in that, Based on the effectiveness status of the comprehensive pH control demand signal and the comprehensive temperature control demand signal, and combined with the preset priority rules of the current fermentation stage, a coordinated control instruction set is generated, including: When both the comprehensive pH control demand signal and the comprehensive temperature control demand signal are valid, the coupled feedback controller first retrieves the preset control priority weights according to the current fermentation stage, and then judges the degree of coupling conflict between the two control actions: if it is a weak conflict or mutually beneficial, it outputs a parallel full-intensity command, so that the two execution units execute the actions at their respective calculated intensities simultaneously; if it is a strong conflict, it executes a step-by-step alternating output strategy, dividing a control cycle into multiple sub-cycles. In each sub-cycle, the high-priority control action is executed first, and the low-priority control demand is attenuated and converted into a feedforward compensation amount and superimposed on the high-priority action command. After the sub-cycle ends, the roles are switched, and this alternation continues until both comprehensive demand signals drop below the dead zone.

10. The intelligent control method for antibacterial agent fermentation production based on pH and temperature coupling feedback as described in claim 9, characterized in that, The control priority weights are preset according to different fermentation stages: during the cell growth stage, temperature control priority is higher than pH control priority. During the bacteriostatin synthesis stage, pH regulation has a higher priority than temperature regulation. During the autolysis phase of the cells, temperature control takes precedence over pH control.

Images