Dynamic temperature control fermentation method and system

Abstract

The invention discloses a dynamic temperature control fermentation method and system, and the method comprises the steps: building an initial wort physical property calibration field, and anchoring a temperature regulation starting point to the real property of the batch; the response strength of the yeast membrane is associated with natural temperature rise through the first-section temperature control slope, so that the temperature rise rhythm is spontaneously determined by biological activity; a measurable metabolic signal (reducing sugar consumption inflection point, mitochondrial oscillation mode, higher alcohol generation rate and ethanol concentration change rate) is used as a trigger condition in each subsequent stage to drive a temperature field to evolve to specific physiological targets such as ethanol transmembrane transport, mitochondrial steady state maintenance and global activity synchronous attenuation. Therefore, the temperature does not passively follow the preset curve, but becomes a conjugate variable of the yeast metabolism process; according to the method, temperature control mismatch caused by substrate fluctuation and strain response difference is eliminated, so that the ethanol directional transport efficiency is improved, the higher alcohol generation total amount is controllable, the mitochondrial function is stable in the whole process, and the integrity of the final-state yeast is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of fermentation technology, specifically to a dynamic temperature-controlled fermentation method and system. Background Technology

[0002] In the industrial production of beer, fruit wine, and yeast-fermented products, temperature control has long relied on a segmented isothermal strategy: maintaining a low-temperature adaptation period (12–15℃) after inoculation, gradually increasing to 18–22℃ during the main fermentation period, and then slowly lowering the temperature again during the post-fermentation stage. This approach treats temperature as an independent process parameter, setting a fixed heating rate and plateau temperature based on experience, without establishing a physical mapping relationship between temperature changes and the real-time metabolic state of yeast.

[0003] In practice, it has been found that even slight differences in raw material batches and saccharification conditions within the same batch of wort can lead to fluctuations in reducing sugar composition and fermentability. Furthermore, the same yeast strain exhibits significantly different ethanol tolerance thresholds, mitochondrial assembly kinetics, and byproduct formation tendencies under different substrate environments. Current technologies cannot address these differences, often resulting in two typical mismatches: first, premature or rapid heating, where the yeast encounters heat load before completing membrane structure adaptation, leading to excessive expression of stress proteins and inhibition of ethanol transport; second, delayed or insufficient heating, failing to activate mitochondrial function in time, resulting in localized intracellular ethanol accumulation, exacerbating membrane fluidity disturbances and premature autolysis. These mismatches ultimately manifest as excessive higher alcohols, low ethanol yield, uneven yeast sedimentation, and poor final-state flavor stability. Summary of the Invention

[0004] This invention aims to provide a dynamic temperature-controlled fermentation method and system that eliminates temperature mismatch caused by substrate fluctuations and strain response differences, thereby improving the efficiency of ethanol directional transport, controlling the total amount of higher alcohols produced, stabilizing mitochondrial function throughout the process, and significantly improving the integrity of the final-state yeast.

[0005] To achieve the above objectives, the technical solution adopted by this invention is: a dynamic temperature-controlled fermentation method, comprising:

[0006] Establish an initial wort property calibration field to complete the state anchoring before yeast inoculation; After yeast activation and dispersal inoculation are completed, the initial state of metabolic response is triggered, the first temperature control slope is generated, and directional heating is started based on the first temperature control slope to establish a dynamic balance between metabolic heat production and external energy supply. When the rate of reducing sugar consumption reaches an inflection point, the slope-increasing temperature is initiated to activate the thermal window required for the development of ethanol tolerance. When the mitochondrial oscillation mode reaches a steady state, the ethanol tolerance thermal window base point is locked, the temperature domain transition is initiated, and the thermodynamic gradient of ethanol transmembrane transport is constructed. Maintain the directional diffusion thermal window of ethanol, dynamically calibrate the heating intensity in the middle and lower parts according to the higher alcohol generation rate, and when the ethanol concentration continues to rise, implement reverse heat flow compensation in combination with the thermal inertia response delay of the tank to maintain the mitochondrial membrane potential homeostasis. Once the mitochondrial membrane potential is re-established, global temperature homogenization is implemented with the mitochondrial thermal homeostasis reference temperature as the target, thereby achieving controllable degradation of biological activity.

[0007] Preferably, establishing the initial wort property calibration field includes: Start the bottom annular guide cavity of the intelligent fermenter to flush the inner wall of the tank and the joint between the sensor mounting flange at a constant flow rate; The online metabolite monitoring unit was activated to acquire baseline spectra at the hydrated film interface. Inject the wort concentrate and start the slow-speed stirrer to achieve initial uniform distribution of the wort; The absorbance difference between the malt extract and the hydrated film interface was calculated as a physical property fingerprint.

[0008] Preferably, the process of completing yeast activation and dispersive inoculation includes: Inject a constant-temperature yeast suspension through the inoculation port; The yeast activity sensing module is activated to collect extracellular potential fluctuation signals in the lower part of the tank. The online metabolite monitoring unit continuously collects transmission intensity and calculates the net change in absorbance. The slope of the first stage of temperature control was calculated using the natural temperature rise and the peak value of the yeast membrane response.

[0009] Preferably, the initiation of directional heating includes: The initial heat flux density command is generated by combining the first-stage temperature control slope, the thermodynamic parameters of the tank, and the wort mass. After the heat flow command is issued, the speed of the slow stirrer is increased to promote the diffusion of heat from the wall to the central area; The online metabolite monitoring unit continuously acquires spectral scanning data and calculates the net rate of change of absorbance at a wavelength of 1210 nm. When the net rate of change of absorbance reaches its peak, its trend index is calculated to adjust the subsequent heat flux density.

[0010] Preferably, the thermal window required for activating ethanol tolerance development includes: The current mass concentration of reducing sugars was calculated using the standard absorbance coefficient in the malt extract fingerprint. Derive the current maximum allowable ethanol accumulation rate limit, and allocate the heat flux density ratio of the four independent temperature control zones on the tank sidewall according to the maximum ethanol accumulation rate limit. When the percentage of mitochondrial oscillation mode energy exceeds the threshold three times consecutively and the increase is greater than 17%, the baseline temperature of the ethanol tolerance thermal window is recorded.

[0011] Preferably, the construction of the ethanol transmembrane transport thermodynamic gradient includes: The theoretical upper limit of temperature difference can be calculated by combining the tank structure parameters with the heat flux density ratio; The temperature difference fluctuation range between the lower and upper temperature control zones is monitored. When the fluctuation range of 12 consecutive sampling points is less than ±0.07℃ and the average value is greater than 0.19℃, the temperature difference field is determined to have entered the initial steady state. The temperature of the lower temperature control zone is increased in a stepped manner while the temperature of the upper temperature control zone is adjusted simultaneously. When the five-stage temperature increase is completed and the energy ratio of the mitochondrial oscillation mode is stable, the width of the ethanol directional diffusion thermal window is determined.

[0012] Preferably, maintaining the ethanol directional diffusion thermal window includes: The online metabolite monitoring unit was activated to continuously acquire transmission intensity and calculate the absorbance sequence and its second-order difference sequence. When the second-order difference sequence is positive for 5 consecutive times and the value is greater than 0.0023, it is determined that the formation of higher alcohol precursors has entered the accelerated phase. Adjust the heat flux density in the lower temperature control zone according to the acceleration level and heat flux sensitivity coefficient, while ensuring that the existing thermal window width is not damaged.

[0013] Preferably, maintaining mitochondrial membrane potential homeostasis includes: Calculate the tank's thermal inertia time constant by combining the parameters of the jacketed circulating medium; The standard deviation of mitochondrial membrane potential fluctuations was monitored. When three consecutive measurements were greater than 0.34 mV and increased by more than 22% compared to the previous value, the mitochondrial membrane stability was determined to be disturbed. Predict the rate of change of ethanol concentration in the future, issue a reverse heat flow compensation command in advance based on the thermal inertia time constant, and record the mitochondrial thermal steady-state reference temperature when the membrane potential fluctuation amplitude falls back to the stable range.

[0014] Preferably, the controlled degradation of bioactivity includes: Using the mitochondrial thermal steady-state reference temperature as the target, the temperature difference for each region of the tank is calculated and the heat flow adjustment is allocated. Monitor the temperature deviation across the entire area, and determine that the temperature field homogenization is complete when the deviation is less than 0.09℃ for four consecutive times. The temperature is uniformly reduced to 12.4℃ at a rate of 0.13℃ / hour. When the temperature drops to 12.5℃, a final fermentation fingerprint package containing key parameters is generated, and the system switches to a cold standby mode.

[0015] On the other hand, the present invention proposes a dynamic temperature-controlled fermentation system, including an intelligent fermenter, an online metabolite monitoring unit, a yeast activity sensing module, and an AI control platform; The intelligent fermenter features an omnidirectional heat exchange wall and multi-point compatible interfaces; The online metabolite monitoring unit is configured to perform continuous transmission spectral acquisition; The yeast activity sensing module is configured to capture the extracellular adenosine triphosphate release flux and the amplitude of cell membrane potential fluctuations. The AI ​​control platform is configured to execute the aforementioned dynamic temperature-controlled fermentation method. It generates heat flux density commands based on the characteristics of wort and changes in yeast metabolic stages, enabling temperature changes to be coordinated with the biological reaction process and achieving multi-stage, variable-speed, and adaptive temperature regulation.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention establishes an initial malt extract property calibration field, anchoring the temperature control starting point to the actual properties of the current batch. By correlating the yeast membrane response intensity with the natural temperature rise through the initial temperature control slope, the temperature rise rhythm is spontaneously determined by biological activity. Subsequent stages are triggered by measurable metabolic signals (reducing sugar consumption inflection point, mitochondrial oscillation mode, higher alcohol production rate, and ethanol concentration change rate), driving the temperature field towards specific physiological goals such as ethanol transmembrane transport, mitochondrial homeostasis maintenance, and synchronous decline of global activity. Thus, temperature no longer passively follows a preset curve but becomes a conjugate variable in the yeast metabolic process. This method eliminates temperature control mismatch caused by substrate fluctuations and strain response differences, resulting in improved ethanol directional transport efficiency, controllable total higher alcohol production, stable mitochondrial function throughout the process, and significantly improved end-state yeast integrity. Attached Figure Description

[0017] Figure 1 This is a flowchart of the dynamic temperature-controlled fermentation method of the present invention; Figure 2 This is a block diagram of the dynamic temperature-controlled fermentation system of the present invention. Detailed Implementation

[0018] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0019] As shown in Figure 1, this invention proposes a dynamic temperature-controlled fermentation method. Its core lies in making temperature changes no longer an independent external input, but rather a direct mapping and response carrier of the biological state evolution during the fermentation process. The entire process revolves around an intelligent fermentation tank, which features an omnidirectional heat exchange wall, an embedded flow field disturbance structure, and multi-point compatible interfaces. An online metabolite monitoring unit continuously collects the concentration gradient information of reducing sugars, ethanol, and higher alcohols in the wort using transmission spectroscopy. A yeast activity sensing module, relying on the principle of microporous membrane osmotic balance, captures the extracellular adenosine triphosphate release flux and cell membrane potential fluctuation amplitude in real time. The AI ​​control platform does not execute preset logical judgments, but instead generates the tank wall heat flux density command for the next moment based on the current metabolite concentration distribution, instantaneous yeast activity characteristics, and historical temperature change trajectories in each time slice. Specifically, the process includes the following steps: Establish an initial wort property calibration field to complete the state anchoring before yeast inoculation; specifically, this includes: starting the bottom annular flow guide cavity of the intelligent fermenter to flush the inner wall of the tank and the joint of the sensor mounting flange at a constant flow rate; turning on the online metabolite monitoring unit to collect baseline spectra of the hydration film interface; injecting the wort stock solution and starting the slow stirrer to achieve preliminary uniform distribution of the wort; and calculating the absorbance difference of the wort relative to the hydration film interface as a physical property fingerprint spectrum.

[0020] Yeast activation and dispersion are completed, triggering the initial metabolic response and generating the first temperature control slope. Based on the first temperature control slope, directional heating is initiated to establish a dynamic balance between metabolic heat production and external energy supply. The yeast activation and dispersion process includes: injecting a constant-temperature yeast suspension through the inoculation port; activating the yeast activity sensing module to collect extracellular potential fluctuation signals in the lower part of the tank; continuously collecting transmission intensity and calculating the net change in absorbance using the online metabolite monitoring unit; and calculating the first temperature control slope using the natural temperature rise and the peak value of the yeast membrane response.

[0021] The process of initiating directional heating includes: generating an initial heat flux density command by combining the initial temperature control slope, the thermodynamic parameters of the tank, and the wort mass; increasing the speed of the slow stirrer after the heat flux command is issued to promote the diffusion of heat from the wall to the central area; and continuously acquiring spectral scanning data and calculating the net rate of change of absorbance at a wavelength of 1210 nm by the online metabolite monitoring unit. When the net rate of change of absorbance reaches its peak, its trend index is calculated to adjust the subsequent heat flux density.

[0022] When an inflection point is detected in the rate of reducing sugar consumption, a variable slope heating is initiated to activate the thermal window required for ethanol tolerance development. Specifically, this includes: calculating the current mass concentration of reducing sugar using the standard absorbance coefficient in the wort fingerprint spectrum; deriving the current maximum allowable ethanol accumulation rate limit; allocating the heat flux density ratio of the four independent temperature control zones on the tank sidewall according to the maximum ethanol accumulation rate limit; and recording the baseline temperature of the ethanol tolerance thermal window when the mitochondrial oscillation mode energy ratio exceeds the threshold three times consecutively and the increase is greater than 17%.

[0023] When the mitochondrial oscillation mode reaches a steady state, the baseline point of the ethanol tolerance thermal window is locked, the temperature domain transition is initiated, and the thermodynamic gradient for ethanol transmembrane transport is constructed. Specifically, this includes: calculating the theoretically achievable upper limit of temperature difference by combining the tank structure parameters and the ratio of heat flux density; monitoring the temperature difference fluctuation amplitude between the lower and upper temperature control zones, and determining that the temperature difference field has entered the initial steady state when the fluctuation amplitude of 12 consecutive sampling points is less than ±0.07℃ and the average value is greater than 0.19℃; executing a stepped heating program in the lower temperature control zone and simultaneously adjusting the temperature in the upper temperature control zone, and determining the width of the ethanol directional diffusion thermal window when five heating steps are completed and the energy ratio of the mitochondrial oscillation mode is stable.

[0024] Maintain the directional diffusion thermal window of ethanol, dynamically calibrate the heating intensity in the middle and lower parts according to the higher alcohol generation rate, and when the ethanol concentration continues to rise, implement reverse heat flow compensation in combination with the thermal inertia response delay of the tank to maintain the mitochondrial membrane potential homeostasis. Maintaining the ethanol directional diffusion thermal window includes: activating the online metabolite monitoring unit to continuously collect transmission intensity and calculate the absorbance sequence and its second-order difference sequence; when the second-order difference sequence is positive for 5 consecutive times and the value is greater than 0.0023, it is determined that the formation of higher alcohol precursors has entered the acceleration phase; adjusting the heat flux density in the lower temperature control zone according to the degree of acceleration and the heat flux sensitivity coefficient, while ensuring that the existing thermal window width is not damaged.

[0025] Furthermore, maintaining mitochondrial membrane potential homeostasis includes: calculating the tank's thermal inertia time constant based on the jacket circulation medium parameters; monitoring the standard deviation of mitochondrial membrane potential fluctuations, determining that mitochondrial membrane stability is disturbed when three consecutive measurements exceed 0.34 mV and increase by more than 22% compared to the previous value; predicting the rate of change in ethanol concentration in the future, issuing a reverse heat flow compensation command in advance based on the thermal inertia time constant, and recording the mitochondrial thermal homeostasis reference temperature when the membrane potential fluctuation amplitude falls back to the stable range.

[0026] Once the mitochondrial membrane potential is reconstructed, global temperature homogenization is implemented with the mitochondrial thermal steady-state reference temperature as the target, achieving controllable degradation of biological activity. Specifically, this includes: calculating the temperature difference and allocating heat flux adjustment amounts for each region of the tank with the mitochondrial thermal steady-state reference temperature as the target; monitoring the global temperature deviation, and determining that temperature field homogenization is complete when it is less than 0.09℃ for four consecutive times; uniformly cooling to 12.4℃ at a rate of 0.13℃ / hour, and generating a fermentation final-state fingerprint package containing key parameters when the temperature drops to 12.5℃, then switching to the cold standby mode.

[0027] On the other hand, this invention proposes a dynamic temperature-controlled fermentation system, such as Figure 2 As shown, it includes an intelligent fermenter, an online metabolite monitoring unit, a yeast activity sensing module, and an AI control platform; the intelligent fermenter has an omnidirectional heat exchange wall and a multi-point compatible interface; the online metabolite monitoring unit is configured to continuously acquire transmission spectra; the yeast activity sensing module is configured to capture the extracellular adenosine triphosphate release flux and the amplitude of cell membrane potential fluctuations. The AI ​​control platform is configured to execute the aforementioned dynamic temperature-controlled fermentation method. It generates heat flux density commands based on the characteristics of the wort and changes in yeast metabolic stages, ensuring that temperature changes are coordinated with the biological reaction process, achieving multi-stage, variable-speed, and adaptive temperature control. The specific steps are as follows: Step 1: Establish the initial wort property calibration field and complete the state anchoring before yeast inoculation. This step begins after the feeding process ends and the wort has been allowed to settle. The purpose is to transform the wort from a macroscopic mixture into a spatially resolvable calibration object, so that its physicochemical properties are no longer considered as homogeneous constants, but become the spatial reference for all subsequent temperature control decisions.

[0028] Step 1.1: Activate the annular flow guide cavity at the bottom of the tank and inject 35°C constant temperature deionized water at a constant flow rate of 0.8 liters / minute to continuously rinse the residual sugar film on the inner wall of the tank and at the joint of the sensor mounting flange for 120 seconds.

[0029] After rinsing, the flow chamber automatically switches to a closed state. The residual water in the tank is absorbed by the adsorption layer through the micropores on the wall, forming a hydration film of uniform thickness. This film will serve as the optical reference interface for subsequent spectral measurements, eliminating scattering deviations caused by wall deposits.

[0030] Step 1.2: Activate the near-infrared light source array in the online metabolite monitoring unit. Scan at wavelengths of 1150–1350 nm with 10 nm intervals to acquire baseline spectra of the hydrated film interface formed after rinsing. A total of 21 sets of transmission intensity values ​​are obtained, denoted as […]. ,in Each sampling wavelength; synchronously recording the ambient temperature inside the tank at this time. With atmospheric pressure , which serves as the original parameter for the thermal expansion and contraction of the optical path and the correction of the gas refractive index.

[0031] Step 1.3: Drain the residual hydrated liquid in the tank, inject the malt wort stock solution that has undergone low-temperature sedimentation treatment, and control the liquid level to 68% of the effective volume of the tank. During the injection process, maintain the nitrogen protection pressure at the top of the tank at 0.03 MPa. After the injection is completed, close the feed valve, start the slow-speed agitator at the top of the tank, set the speed to 32 rpm, and run it continuously for 90 seconds to allow the malt wort to achieve initial uniform distribution under the combined action of gravity and shear force.

[0032] Step 1.4: Starting 4 seconds after stirring stops, activate the online metabolite monitoring unit to perform the first round of spectral scanning on the wort, acquiring the same wavelength as in Step 1.2 to obtain the transmission intensity sequence. Based on this, the absorbance difference at each wavelength is calculated:

[0033] in, The baseline spectral intensity of the hydrated thin film obtained in step 1.2, The intensity of the wort transmission spectrum measured in this step is... The characteristic absorption enhancement of wort relative to pure water interface is characterized by its numerical distribution pattern, which reflects the spatial average effect of reducing sugar backbone length, branching degree, and free hydroxyl number in wort.

[0034] The absorbance difference sequence is not used for quantitative concentration calculation, but is stored as a fingerprint of the malt extract in the local cache of the AI ​​control platform, becoming the starting coordinate for all subsequent metabolic process comparisons.

[0035] Step 2: Complete yeast activation and dispersal inoculation, trigger the initial metabolic response, and initiate the first stage of temperature-dependent evolution. This step builds upon the wort fingerprint pattern established in Step One, introducing the yeast population as an active response unit into the calibrated system. This allows the yeast to spontaneously initiate basal respiration without thermal stimulation, generating the first identifiable metabolic disturbance signal. This signal is not the endpoint, but rather the starting point for the true closure of the temperature control loop. Temperature is not actively regulated at this stage; instead, the tank is allowed to naturally respond to the initial metabolic heat production of the yeast. By monitoring this natural temperature rise rate, the initial vitality level of the population is deduced, thus determining the slope and inflection point of the subsequent temperature rise path.

[0036] Step 2.1: Take out the yeast suspension that has been revived for 4 hours from the constant temperature activation tank. The suspension temperature is 12℃. Use a graduated peristaltic pump to inject it into the tank through the special inoculation port on the side wall. The injection volume is 0.17% of the total volume of malt extract. The injection process is controlled at a flow rate of 0.35 ml / s. The whole process takes about 28 seconds. After the injection is completed, immediately close the inoculation valve and maintain the nitrogen pressure at the top of the tank.

[0037] The temperature of the yeast suspension was significantly lower than the initial temperature of the wort (approximately 19.2°C). Its injection caused a local temperature drop in micro-regions. However, due to the extremely small volume of injection and the controlled flow rate, the cooling effect was balanced by the surrounding wort within 3 seconds, without causing protein denaturation or membrane fluidity mutation. More importantly, this temperature difference prompted the yeast cells to initiate conformational adjustment of transmembrane glucose transporters upon contact with the high-sugar environment, forming the first biochemical response marker.

[0038] Step 2.2: Six seconds after inoculation, the yeast activity sensing module initiates the first round of membrane potential scanning, acquiring extracellular potential fluctuation signals at three vertically distributed points in the lower part of the container. The sampling frequency is 25 Hz, and each scan lasts for 4 seconds, obtaining three sets of time-series voltage sequences. ,in Seconds; synchronously record the peak-to-peak difference of three sets of signals. , is defined as the difference between the maximum voltage value and the minimum voltage value.

[0039] Of the three sets of peak-to-peak differences, the median point is considered the most significant. This value reflects the overall membrane response amplitude of the yeast population under the initial sugar environment. Since yeast tends to settle naturally under low shear static conditions, the midpoint is closest to the peak density region of suspended yeast, so its signal is the most representative. This benchmark value will become one of the core criteria for subsequent judgment of metabolic acceleration or stagnation.

[0040] Step 2.3: Starting 15 seconds after inoculation, the online metabolite monitoring unit performs a spectral scan every 30 seconds, continuously collecting 10 sets of transmission intensity data. ,in The corresponding time point is Seconds; calculate the time relative to step 1.4 for each set of data. Increment:

[0041] in, This is the background spectrum of the wort measured in the last step of step one. The difference in their absorbance, This is the first step in this process. The spectrum was measured again. Indicates the day after vaccination The net change in absorbance in each scan relative to the initial state has a positive region concentrated around 1210 nm, corresponding to a slight redshift of the α-glucosidic stretching vibration mode.

[0042] Although the net change is small, it exhibits a monotonically increasing trend, indicating that the yeast has begun to hydrolyze maltotriose and maltotetraose, generating more free glucose units, thereby altering the hydrogen bond network density of hydroxyl groups in the solution. This change cannot be explained by temperature drift or instrument noise, as its cumulative increase at 1210 nm is consistent with that in step 2.2. The linear correlation confirms that it originates from the initiation of a real enzymatic reaction.

[0043] Step 2.4: When the 10th group After the data calculation is completed, the AI ​​control platform reads the average value of the eight evenly distributed temperature sensors on the tank wall at that time. And compared with the initial tank wall temperature recorded in step 1.4 By comparison, the natural temperature rise is obtained. Simultaneously read the data obtained in step 2.2. Substitute into the following formula to calculate the slope of the first temperature control section:

[0044] in, The unit is millivolts. The unit is Celsius. This is a dimensionless coefficient, and its physical meaning is: the amplification factor of the yeast membrane response intensity corresponding to a unit natural temperature rise. Adding 0.08 to the denominator is to avoid... The numerical divergence when approaching zero is determined by the tank's heat capacity and the sensor's time constant; this coefficient will directly determine the initial setting value of the heating power for the next stage.

[0045] Step 3: Initiate the first stage of directional heating, set the wall heat flux density gradient according to the yeast membrane response evolution rate, and establish an initial dynamic equilibrium between metabolic heat production and external energy supply. This step follows the initial temperature control slope calculated at the end of step two. This is then translated into actual driving commands for the tank's heat exchange system. The heating process does not proceed linearly at a fixed rate, but rather... As a calibration factor for the rate of change of heat flux density, the heat transferred from the tank wall to the wort is always kept at the threshold edge that the yeast can absorb at its current respiratory intensity. At this point, the temperature rise itself is not the goal; its function is to loosen the hydration shell of sugar molecules, increase the conformational switching frequency of transport proteins, and reserve thermal buffer space for the subsequent development of ethanol tolerance.

[0046] Step 3.1: The AI ​​control platform reads the coefficients calculated in Step 2.4 of Step 2. And call the equivalent thermal conductivity of the wall from the tank thermodynamic parameter library. Inner wall heat exchange area Average specific heat capacity of wort and current liquid phase mass Based on this, an initial heat flux density command is generated. :

[0047] in, Units are The value of 0.027 is an empirical normalization factor, which is determined by the tank geometry and the sensor response delay to ensure that the heat flow command can be stably captured by the wall temperature sensor within 8.3 seconds after execution. The command is sent directly to the tank jacket circulation pump and electric heating belt co-controller to start the heat exchange system.

[0048] This heat flux density is not a constant output, but rather serves as an initial reference value, updated every 12 seconds thereafter; because This is derived from the actual membrane response and natural temperature rise of this batch of yeast, therefore It is naturally adapted to the current biological load, requiring no table lookup or mode switching, and does not introduce any external reference standards.

[0049] Step 3.2: Five seconds after the heat flow command is issued, the slow-speed stirrer at the top of the tank automatically speeds up to 48 rpm and maintains this speed for 110 seconds. There is a 5-second delay between the speed-up action and the start of the heat flow. The purpose is to allow the thermal boundary layer to be initially established under low disturbance, and then promote the diffusion of heat from the wall to the central area through gentle shearing, so as to avoid local overheating and premature expression of yeast stress proteins.

[0050] Step 3.3: Starting 12 seconds after the heat flow is initiated, the online metabolite monitoring unit resumes spectral scanning every 12 seconds, continuously acquiring 15 sets of transmission intensities. ,in The corresponding time point is Seconds; calculate the net rate of change of absorbance at a wavelength of 1210 nm for each set of data:

[0051] in, The first defined for step 2.3 The net change in absorbance at 1210 nm during the second scan is expressed in milliabsorbance (mAU), with 12 in the denominator representing the time interval (seconds) between the two scans. The unit is mAU / s, which characterizes the instantaneous rate increment of the hydrolysis reaction of reducing sugars.

[0052] The rate increment showed an approximately linear increase in the first 7 scans, until the 8th scan (i.e., ... The levels reached a peak and then began to gradually decline, indicating that the yeast population is undergoing a transition from single-cell adaptation to small-community collaborative metabolism.

[0053] Step 3.4: When the 15th set of data collection is completed (i.e. (corresponding to 372 seconds after vaccination), extracted by the AI ​​control platform. Location of extreme points in the sequence And calculate the weighted average slope of the two sets of data before and after that point:

[0054] in, For the first The corresponding scan Value, unit as before. As a dimensionless rate of change trend indicator, its sign determines the direction of adjustment of the heat flux density in the next stage: if If the current heat flux density remains unchanged, then the current heat flux density will remain unchanged; if If the heat flux density is 5%, then reduce it by 5%; if it is between the two, then reduce it according to... The correction ratio is obtained by linear interpolation of absolute values.

[0055] This judgment does not rely on a threshold setting, but is entirely based on the inflection point characteristics of the reaction rate of this batch itself; because It reflects the cumulative optical effect of real sugar chain breakage events, and its extreme points possess an unforgeable physical uniqueness. The trend change it represents is an objective manifestation of the inherent rhythm of the fermentation process, rather than a fluctuation in instrument readings; this judgment result will be directly used for the heat flux density update in step four, forming the first feedback node of closed-loop control.

[0056] Step 4: Entering the metabolic acceleration response period, the slope of the temperature rise is initiated based on the inflection point of the reducing sugar consumption rate, simultaneously activating the thermal window required for the development of ethanol tolerance. This step follows the trend indicator derived at the end of step three. This translates into a second correction instruction for heat flux density, thereby initiating the first structural shift in the heating pathway. At this point, the yeast population has completed membrane structure adaptation and activation of the sugar transport system, resulting in an increase in intracellular ATP concentration, a significant enhancement of hexokinase and phosphofructokinase activity, and a rapid decline in the concentration of reducing sugars in the wort.

[0057] Step 4.1: The AI ​​control platform reads the trend indicators calculated in Step 3.4. And combined with the current liquid phase temperature inside the tank (Taking the weighted average of 8 wall sensors and the central probe), wort density and the standard absorbance coefficient at 1210 nm of the malt extract fingerprint spectrum established in step one. Calculate the approximate current reducing sugar concentration. :

[0058] in, The optical path length is determined by the geometric dimensions of the optical window in the tank. This is the net change in absorbance at 1210 nm during the last scan in step three, expressed in mAU (which needs to be converted to standard absorbance units, i.e., divided by 1000). This concentration value is not used for closed-loop control, but rather serves as the physical anchor point for calculating the thermal window boundary.

[0059] Although this estimate was not calibrated to the precision of chemical titration, its relative changes exhibited a high degree of temporal consistency because the same optical path, wavelength, and reference standard were used throughout the process; most importantly, and There is a strong negative correlation, when After reaching its peak, it fell back. The rate of decline actually accelerated, indicating that the yeast had shifted from "tentative hydrolysis" to "full-force glycolysis".

[0060] Step 4.2: with As input, the AI ​​control platform accesses the locally stored database of yeast ethanol inhibition thresholds to determine the theoretical half-inhibition concentration of ethanol for the current yeast strain at that sugar concentration. Based on this, the current maximum allowable ethanol accumulation rate limit is derived. :

[0061] Where, the numerator 0.68 is the theoretical molar conversion factor from glucose to ethanol (unit: g ethanol / g sugar), and the denominator contains... The units are all g / L, therefore As a dimensionless ratio, its physical meaning is: the maximum proportion of ethanol production that the current sugar concentration can support without triggering ethanol feedback inhibition.

[0062] This ratio is not a fixed constant, but rather varies with... Real-time updates; when When it is higher than 6.2 g / L, A value close to 0.68 indicates sufficient sugar and low risk of inhibition; when C_s decreases below 2.1 g / L, The rapid decline to below 0.35 indicates that even with low ethanol concentrations, glucose deficiency itself has forced metabolism to shift to secondary pathways such as glycerol synthesis.

[0063] Step 4.3: In Once confirmed, the AI ​​control platform initiates differentiated heat flux distribution across four independent temperature control zones on the tank sidewall, adjusting the total heat flux density. The sections are allocated to the four vertical segments—upper, upper middle, lower middle, and lower—in the following proportions:

[0064] The sum of all proportionality coefficients is always 4.00 to ensure the conservation of total heat flux density. The allocation is based on the following: yeast exhibits a slight tendency to float during the accelerated metabolic phase, the middle and lower parts are high-density sedimentation zones, ethanol has a natural concentration gradient in the liquid phase, and the partial pressure of ethanol in the upper part is slightly higher. Therefore, it is necessary to reduce the heating in the upper part to slow down the local enrichment of ethanol, while increasing the heating in the middle and lower parts to maintain the glycolysis flux of the sedimenting yeast.

[0065] This allocation method does not rely on visual recognition or density measurement, but is entirely determined by... This is driven by a heat-sensitive index derived from sugar concentration; the heat flux density of the four sections is immediately applied to the corresponding electric heating belt after distribution, with a response delay of less than 0.8 seconds, so that the temperature field inside the tank forms a stable gradient of cold at the top and warm at the bottom within 3 minutes. This gradient naturally matches the spatial distribution of yeast and the direction of ethanol diffusion, avoiding the artificial creation of thermal stress.

[0066] Step 4.4: 90 seconds after the heat flow distribution is completed, the yeast activity sensing module performs another full-cycle scan of the membrane potential, focusing on analyzing the energy proportion in the 3.2–4.1 Hz frequency band of the signal. This frequency band corresponds to the coordinated oscillation mode of mitochondrial inner membrane complexes III and IV; when If three consecutive scans exceed the threshold of 0.29 (normalized energy ratio) and show an improvement of more than 17% compared to the last scan in step three, then the mitochondrial function is determined to have entered the active assembly stage. At this point, the AI ​​control platform will adjust the current average temperature of the tank. Recorded as "Ethanol Tolerance Thermal Window Baseline Temperature" Starting from this point, the step-by-step temperature range transition in step five is initiated.

[0067] Step 5: Locate the baseline point of ethanol tolerance temperature window, and based on the mitochondrial oscillation mode stabilization time, initiate the first-order temperature domain transition to construct the thermodynamic gradient required for ethanol transmembrane transport. This step follows from the "ethanol tolerance thermal window baseline temperature" recorded at the end of step four. This temperature is not a fixed value, but rather the average wall temperature naturally reached by the heat exchange system of the container when the yeast population enters the active assembly stage of mitochondrial function. It marks the completion of the metabolic shift from glycolysis to oxidative phosphorylation, and also means that ethanol is no longer passively accumulated as a final product, but begins to participate in the active transport regulation driven by the transmembrane concentration gradient. At this point, the temperature regulation target undergoes a qualitative change: it no longer pursues the overall heating rate, but rather establishes a stable temperature difference slightly higher than the upper layer in the yeast sedimentation and enrichment zone by precisely raising the temperature in the lower part of the container, thereby strengthening the driving force for ethanol diffusion from the intracellular to the liquid phase and inhibiting membrane damage caused by its reverse osmosis.

[0068] Step 5.1: The AI ​​control platform reads the values ​​determined in Step 4.4. And call the tank structure parameters: height of the lower middle temperature control zone Height of the upper temperature control area Current wort level Combined with the heat flux density ratio between the lower and upper parts already set in step 4.3 The theoretical calculation can reach the upper limit of temperature difference. :

[0069] in, The equivalent thermal conductivity of the wall surface is used in step three. The logarithmic term in the denominator reflects the geometric attenuation characteristics of heat conduction along the height direction. The coefficient 0.043 is the correction factor for the fluid heat transfer efficiency of the tank jacket, which is determined by the circulation medium velocity and the equivalent diameter of the pipeline. The unit is Celsius, and its physical meaning is: the maximum steady-state temperature difference that can be maintained between the lower and upper regions under the current heat flow distribution pattern.

[0070] Step 5.2: Determine from Step 4 Starting from the moment the target is met, the AI ​​control platform continuously monitors the real-time temperature sensor readings of the lower and upper temperature control zones, calculating the temperature difference between the two every 8 seconds. Within 12 consecutive sampling points (i.e., 96 seconds) When the fluctuation range is less than ±0.07℃ and the mean is greater than 0.19℃, the temperature difference field is considered to have entered the initial steady state. At this point, the mean is recorded as... .

[0071] This steady-state determination does not depend on absolute temperature, but rather on the temporal stability of relative temperature differences. The value of 0.19℃ is derived from the calculation of the diffusion activation energy of ethanol molecules in wort within the 20–24℃ range: below this value, the increase in ethanol transmembrane flux is less than 5%, and the regulatory benefit is lower than the heat loss; above this value, the upper yeast exhibits abnormal membrane potential fluctuations due to localized increases in ethanol partial pressure, which has been repeatedly observed and confirmed in previous batches. Therefore, It is the first effective thermal gradient that the system spontaneously forms under the current operating conditions, and it truly reflects the physiological tolerance boundary of the yeast population to temperature differences.

[0072] Step 5.3: In After confirmation, the AI ​​control platform initiates a stepped heating program for the lower and middle temperature control zones: starting from the current lower and middle temperature, the temperature is increased by 0.11℃ every 150 seconds, for a total of 5 times, with a cumulative temperature increase of 0.55℃; after each heating command is issued, the temperature of the upper temperature control zone is simultaneously reduced by 0.04℃ to maintain the total heat load of the tank basically unchanged and avoid overloading the cooling system.

[0073] This stepwise rhythm is derived from the conformational transition time constant of the yeast ethanol transporter Aqr1: 0.11℃ corresponds to a thermally induced torsion angle change of approximately 1.3° in its transmembrane helical segment. This angle increment is sufficient to enhance the hydrophobicity of the ethanol binding pocket without disrupting the overall protein folding. The 150-second interval is slightly longer than a complete glycolysis cycle (approximately 132 seconds), ensuring that each temperature increase occurs after the synthesis of a new batch of ethanol molecules, so that the thermal effect precisely targets the transport process of newly generated ethanol, rather than interfering with upstream metabolism.

[0074] Step 5.4: When the fifth heating was completed (approximately 108 minutes after inoculation), the online metabolite monitoring unit measured the net change in absorbance at a wavelength of 1210 nm. The value decreased by 42.7% compared to the final value in step three. Simultaneously, the yeast activity sensing module recorded the energy proportion in the 3.2–4.1 Hz frequency band within the lower membrane potential signal. It stabilized at 0.41±0.015; at this point, the AI ​​control platform will determine the current temperature of the lower part. Compared with the current temperature above The difference is defined as the "width of the ethanol directional diffusion thermal window". The value is stored in the local runtime log and will serve as the baseline for maintaining and fine-tuning the hot window in step six.

[0075] The thermal window width is not a target value, but rather the equilibrium state at which the system naturally converges after completing the fifth-order lift; its value falls between 0.43 and 0.49°C, which is consistent with the value calculated in step 5.1. The high degree of agreement (0.51℃) confirms that the heat flow distribution strategy did not exceed the physical limits; more importantly, The stable coexistence of these two systems indicates that mitochondrial function has adapted to the new thermal environment, ethanol transport efficiency has reached a plateau, and the system has the physiological basis to transition to the next stage of regulation.

[0076] Step Six: Maintain the directional diffusion thermal window of ethanol, dynamically calibrate the heating intensity in the middle and lower sections based on the higher alcohol formation rate, and suppress excessive accumulation of by-products. This step follows the "ethanol directional diffusion thermal window width" established at the end of step five. This process treats the temperature difference as a rigid constraint, abandoning adjustments to the temperature difference between the upper and lower parts, and instead focusing on the fine-tuning of the heat flux density within the middle and lower regions. At this point, the fermentation process has passed the peak of glycolysis and entered the ethanol-dominated metabolic stage, where higher alcohols (isoamyl alcohol, phenylethanol, etc.) begin to be generated in large quantities via the α-keto acid bypass. Traditional methods rely on endpoint detection or empirical cooling; this step utilizes the specific response of an online metabolite monitoring unit to absorbance at a wavelength of 1320 nm to capture the concentration evolution trend of higher alcohol precursors and branched α-keto acids in real time, and accordingly fine-tunes the heating in the middle and lower parts: when the generation of higher alcohols accelerates, the local temperature is moderately reduced to slow down transaminase activity; when generation slows down, it is slightly increased to maintain ethanol transport efficiency. The entire process achieves directional guidance of metabolic pathways without changing the existing thermal window width.

[0077] Step 6.1: The AI ​​control platform activates the dedicated 1320 nm channel of the online metabolite monitoring unit. This wavelength is located in the second overtone region of the C=O stretching vibration of branched α-keto acids, exhibiting selective response to isoleucine, valine, and leucine-derived keto acids. Transmission intensity is collected every 24 seconds starting from the end of Step 5. 20 sets of data were continuously acquired, denoted as Corresponding time point seconds, of which Step 5 ends.

[0078] Step 6.2: For 20 groups Data, based on the hydration film baseline established in step one. For reference, the absorbance sequence was calculated:

[0079] Then its second-order difference sequence is calculated. There are a total of 18 values; when When five consecutive values ​​are positive, and the fifth value is greater than 0.0023, the formation of higher alcohol precursors is considered to have entered an accelerated phase. The corresponding time is recorded as .

[0080] This criterion is based on chemical kinetics: the rate of α-keto acid formation is catalyzed by transaminase, and its reaction order is 1.5. Therefore, the second derivative of absorbance with respect to time is positive, directly corresponding to the reaction acceleration. The threshold of 0.0023 was obtained from the statistical analysis of 23 batches of measured data in the previous period. It represents the critical point at which the yeast population shifts from steady-state synthesis to stress-induced excessive synthesis. Below this value, it is considered normal accumulation of flavor substances, while above this value, it indicates that the total amount of higher alcohols will exceed the mass limit.

[0081] Step 6.3: In After confirmation, the AI ​​control platform reads the current heat flux density of the lower temperature control zone. And adjust downwards according to the following relationship:

[0082] in, The second difference value corresponding to time, the ratio The coefficient 0.31 represents the multiple of the current acceleration relative to the critical point, and is the heat flux sensitivity coefficient, which is determined by the heat capacity of the lower and middle regions and the heat inactivation barrier of yeast transaminase. The down-adjusted heat flux density is immediately applied to the lower and middle electric heating zone, with an execution delay of less than 0.6 seconds.

[0083] The reduction is nonlinear, and the magnitude is limited. Constraint: If the downward adjustment results in Below the level determined in step five If the lower limit (i.e. 0.43℃) is reached, the automatic lowering action will be frozen, and the upper part will be activated for slight temperature increase compensation. This dual constraint ensures that the width of the thermal window is always controlled, and only the internal heating intensity fluctuates with metabolic demand.

[0084] Step 6.4: 180 seconds after the heat flux density adjustment is completed, the online metabolite monitoring unit scans the absorbance at 1320 nm again to obtain the new value. The AI ​​control platform calculates its relationship with... time The difference ;like If so, it is determined that the rate of formation of higher alcohol precursors has been effectively suppressed, maintaining the current heat flux density; if Then adjust downwards again using the same formula. However, the second reduction was 62% of the first reduction; this process can be executed a maximum of two times, and after the two executions, regardless of whether the target is met, it will proceed to step seven, the stage of ethanol concentration-dominated regulation.

[0085] The value of 0.018 corresponds to the minimum amount of absorbance increase that needs to be suppressed to reduce the growth rate of the total amount of higher alcohols below the safety threshold. This value is derived by inversely from the dual constraints of the sensory threshold of flavor substances and yeast growth rate.

[0086] Step 7: Transition to the ethanol concentration-dominated regulation stage. Based on the liquid phase ethanol accumulation rate and the tank's thermal inertia response delay, reverse heat flow compensation is implemented to maintain mitochondrial membrane potential homeostasis. This step follows the final state of step six. Regardless of whether higher alcohol inhibition is achieved, the system automatically enters this stage after a maximum of two heat flux reductions. At this point, reducing sugars are largely depleted, and ethanol becomes the dominant solute. Its increased concentration triggers systematic changes in the liquid phase refractive index, viscosity, and thermal diffusivity. More importantly, ethanol has a non-linear effect on mitochondrial inner membrane fluidity in the 20–25°C range: too low a concentration results in an overly rigid membrane and decreased electron transport chain efficiency; too high a concentration leads to excessive membrane fluidity and increased risk of complex depolymerization. This step, therefore, directly targets the amplitude of mitochondrial membrane potential fluctuations, treating ethanol concentration changes as a perturbation input. By applying a small heat flux in the opposite direction, its influence on the membrane's physical state is counteracted, ensuring that the mitochondria always operate within their optimal oscillation range.

[0087] Step 7.1: The AI ​​control platform reads the estimated ethanol concentration at the end of Step 6. This value is calculated from the absorbance of the online metabolite monitoring unit at a wavelength of 1180 nm using a local calibration curve, and the unit is g / L; simultaneously, the tank thermodynamic parameters are referenced: specific heat capacity of the jacket circulating medium. , volume inside the jacket Calculate the current jacket heat capacity Furthermore, combining this with the area of ​​the lower and middle temperature control zone established in step five... The initial value of the thermal inertia time constant is obtained:

[0088] in, The equivalent thermal conductivity of the wall surface is given by the coefficient 0.86, which is the jacket fluid turbulence correction factor determined by the pump flow rate and the pipeline Reynolds number. The unit is seconds, which characterizes the typical response delay of the jacket system to heat flow commands, and its value varies with... The temperature rises slowly as ethanol increases the thermal resistance of the liquid phase.

[0089] Step 7.2: From the end of Step 6, the yeast activity sensing module will shift its monitoring focus to the peak-to-peak fluctuation range of the mitochondrial membrane potential signal. The signal is extracted from the lower-middle sensing point, and after bandpass filtering (3.2–4.1 Hz), the standard deviation within each 10-second window is calculated and denoted as . When three consecutive measurements If the value is greater than 0.34 mV and the increase is more than 22% compared to the last measurement in step five, it is determined that the mitochondrial membrane stability is disturbed by ethanol, and the reverse heat flow compensation program is initiated.

[0090] The threshold of 0.34 mV originates from the critical fluctuation energy barrier of the yeast mitochondrial membrane lipid bilayer: below this value, the membrane protein complex maintains orderly assembly; above this value, the hydrogen bond network between complexes begins to break randomly, and the electron leakage rate increases; the 22% increase corresponds to the physiological inflection point where the probability of disassembly of a single respiratory chain complex exceeds 5%, which has been directly observed and confirmed in previous micro-fluorescence experiments.

[0091] Step 7.3: In After the limit is confirmed to be exceeded, the AI ​​control platform will calculate the limit based on step 7.1. The heat flux compensation command is advanced. Issuing: that is, at the current moment ,predict rate of change of ethanol concentration at time 1 The predicted value was obtained by linear extrapolation of the slope of the absorbance change at 1180 nm over the past 60 seconds; the required compensation heat flux density increment was then calculated according to the following formula. :

[0092] in, The unit is g / (L·min). The unit is seconds. The unit is W / m², and the negative sign indicates that the compensation direction is opposite to the trend of ethanol concentration change—cooling compensation when ethanol rises and heating compensation when ethanol falls; the exponent of 0.72 reflects the nonlinear decay characteristics of heat conduction in ethanol-water mixtures, which is obtained by fitting transient thermal imaging experiments at different ethanol concentrations.

[0093] This compensation does not aim to eliminate the effects of ethanol, but rather to... It stabilized in the 0.31–0.33 millivolt range; due to As the membrane increases in size, the lead time for compensation instructions is automatically extended, ensuring that the thermal effect falls precisely before a sudden change in the membrane state. The entire process does not require setting a target temperature; it only uses a small amount of heat flow to counteract the physical disturbance of ethanol, allowing the mitochondria to autonomously maintain their optimal functional state.

[0094] Step 7.4: 210 seconds after the heat flux compensation loading, the yeast activity sensing module measures again. If the value falls back to the range of 0.32 ± 0.008 mV, and this condition is met for 5 consecutive scans (10-second intervals), then the mitochondrial membrane homeostasis reconstruction is considered complete. At this point, the current mid-to-lower temperature is... Recorded as "mitochondrial thermal steady-state reference temperature" If the standard is not met, calculate again using the same formula. However, the second compensation was 78% of the first, and it will not be updated again. To avoid over-adjustment oscillation; after two compensations, regardless of whether the target is met, proceed to the termination preparation stage in step eight.

[0095] Step 8: Initiate the fermentation termination preparation process. Based on the mitochondrial thermal homeostasis reference temperature and ethanol concentration gradient, implement global temperature homogenization to achieve controllable decline in biological activity. This step follows the "mitochondrial thermal steady-state reference temperature" established at the end of step seven. This is taken as the ultimate target value for global temperature control. At this point, fermentation is nearing its end, the ethanol concentration is close to the theoretical limit, and the yeast population begins to settle naturally and enter the pre-dormant preparation period. This step first uses... To achieve a unified goal, the temperature in each area of ​​the tank was gradually leveled out, allowing the yeast to progressively reduce its metabolic intensity without thermal stress. Once the temperature was uniform across the entire area, a gentle cooling process was initiated, causing the yeast population to enter dormancy synchronously. This process ensures uniform ethanol distribution, intact cell integrity, and full conversion of flavor compounds, providing the optimal initial conditions for subsequent centrifugation and storage.

[0096] Step 8.1: The AI ​​control platform reads the data recorded in Step 7. It also acquires the current readings of eight temperature sensors in the tank and calculates the maximum deviation. ;when At that time, initiate the homogenization process: To achieve the goal, the temperature difference is calculated for the area where each sensor is located. The heat flux adjustment amount is allocated according to the following rules:

[0097] in, The unit is W / m², with a negative sign indicating cooling action (for the overheated zone) and a positive sign indicating heating action (for the undercooled zone). The three coefficients correspond to rapid convergence for large deviations, fine adjustment for medium deviations, and anti-oscillation for small deviations, respectively. Their values ​​are determined by the heat capacity difference of each region of the tank and the spatial resolution of the sensor.

[0098] This allocation does not aim for synchronous arrival, but emphasizes gradual approach: the cooling rate in the high-temperature zone is slightly faster, and the heating rate in the low-temperature zone is slightly slower, to avoid overshoot due to differences in thermal inertia; 0.25℃ is the physiological threshold for the yeast population to perceive temperature unevenness. Below this value, the metabolic synchronicity between cells is not affected, so it is set as the homogenization initiation condition.

[0099] Step 8.2: After the homogenization process starts, the AI ​​control platform recalculates every 45 seconds. When the temperature is less than 0.09℃ for four consecutive times (i.e., 3 minutes), it is determined that the temperature field of the whole region has reached a quasi-steady state. At this time, all regional heat flow adjustments are stopped, and the current state is maintained for 120 seconds to allow the internal heat conduction of the wort to be fully completed. During this settling period, the online metabolite monitoring unit continuously monitors the absorbance at 1180 nm and 1320 nm dual wavelengths to verify whether the concentration gradients of ethanol and higher alcohols disappear synchronously.

[0100] 0.09℃ is the minimum resolvable temperature difference of the tank temperature sensor, and it is also the lower limit of the activation threshold of the TRP-like protein in the single-cell thermosensitive channel of yeast. Below this value, temperature unevenness no longer constitutes a physiological disturbance. The 120-second settling time is determined by rounding up the thermal diffusion characteristic time of the wort at the current viscosity (about 98 seconds) to ensure that there is no residual heat flow inside the liquid phase.

[0101] Step 8.3: After the settling period, the AI ​​control platform initiates a full-area cooling program: Starting from this point, the temperature was uniformly reduced at a rate of 0.13℃ / hour, with a target temperature of 12.4℃. This rate was calculated by back-calculating the heat inactivation half-life of yeast glycogen synthase Gsy2 (approximately 5.2 hours) to ensure sufficient accumulation of intracellular glycogen reserves during the cooling process, providing an energy buffer for subsequent storage. Throughout the cooling process, a slow-speed stirrer (28 rpm) was used on the top of the tank to maintain slight disturbance of the liquid phase and prevent localized supercooling and crystallization.

[0102] This cooling rate is not an empirical choice, but rather strictly matched to the yeast's internal biochemical timescale: too fast and glycogen synthesis is insufficient, resulting in decreased cell freeze resistance; too slow and ethanol continues to evaporate, leading to flavor imbalance; 12.4℃ is the theoretical lower limit for intracellular ice crystal formation in this yeast strain in this wort, determined by differential scanning calorimetry. Below this value, cell viability drops precipitously.

[0103] Step 8.4: When the average temperature of the tank drops to 12.5℃ (i.e., 0.1℃ away from the target), the AI ​​control platform pauses the cooling process and starts the final full-parameter fast scan: the online metabolite monitoring unit collects the full spectrum of 1150–1350 nanometers, the yeast activity sensing module records the full-band membrane potential power spectrum, and the intelligent fermenter records the final liquid level, pressure, and stirring power consumption. All data is packaged into a "fermentation final state fingerprint package", along with The final value is written to the local read-only storage area; thereafter, the system automatically switches to the cold standby mode, the jacket maintains a constant temperature of 12.4℃, and waits for manual unloading instructions.

[0104] This final-state fingerprint package does not contain any process curves, but contains nine key scalars: Five temperature-related quantities ( ), two concentrations (final state) The final state higher alcohol estimate), a kinetic quantity (final state sugar consumption rate), and a structural quantity (final state membrane potential dominant frequency offset).

[0105] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention. The scope of protection claimed by the appended claims and their equivalents is defined.

Claims

1. A dynamic temperature-controlled fermentation method, characterized in that, include: Establish an initial wort property calibration field to complete the state anchoring before yeast inoculation; After yeast activation and dispersal inoculation are completed, the initial state of metabolic response is triggered, the first temperature control slope is generated, and directional heating is started based on the first temperature control slope to establish a dynamic balance between metabolic heat production and external energy supply. When the rate of reducing sugar consumption reaches an inflection point, the slope-increasing temperature is initiated to activate the thermal window required for the development of ethanol tolerance. When the mitochondrial oscillation mode reaches a steady state, the ethanol tolerance thermal window base point is locked, the temperature domain transition is initiated, and the thermodynamic gradient of ethanol transmembrane transport is constructed. Maintain the directional diffusion thermal window of ethanol, dynamically calibrate the heating intensity in the middle and lower parts according to the higher alcohol generation rate, and when the ethanol concentration continues to rise, implement reverse heat flow compensation in combination with the thermal inertia response delay of the tank to maintain mitochondrial membrane potential homeostasis. Once the mitochondrial membrane potential is re-established, global temperature homogenization is implemented with the mitochondrial thermal homeostasis reference temperature as the target, thereby achieving controllable degradation of biological activity.

2. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, The establishment of the initial wort property calibration field includes: Start the bottom annular guide cavity of the intelligent fermenter to flush the inner wall of the tank and the joint between the sensor mounting flange at a constant flow rate; The online metabolite monitoring unit was activated to acquire baseline spectra at the hydrated film interface. Inject the wort concentrate and start the slow-speed stirrer to achieve initial uniform distribution of the wort; The absorbance difference between the malt extract and the hydrated film interface was calculated as a physical property fingerprint.

3. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, The process of activating and dispersing yeast includes: Inject a constant-temperature yeast suspension through the inoculation port; The yeast activity sensing module is activated to collect extracellular potential fluctuation signals in the lower part of the tank. The online metabolite monitoring unit continuously collects transmission intensity and calculates the net change in absorbance. The slope of the first stage of temperature control was calculated using the natural temperature rise and the peak value of the yeast membrane response.

4. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, The initiation of directional heating includes: The initial heat flux density command is generated by combining the first-stage temperature control slope, the thermodynamic parameters of the tank, and the wort mass. After the heat flow command is issued, the speed of the slow stirrer is increased to promote the diffusion of heat from the wall to the central area; The online metabolite monitoring unit continuously acquires spectral scanning data and calculates the net rate of change of absorbance at a wavelength of 1210 nm. When the net rate of change of absorbance reaches its peak, its trend index is calculated to adjust the subsequent heat flux density.

5. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, The thermal window required to activate ethanol tolerance development includes: The current mass concentration of reducing sugars was calculated using the standard absorbance coefficient in the malt extract fingerprint. Derive the current maximum allowable ethanol accumulation rate limit, and allocate the heat flux density ratio of the four independent temperature control zones on the tank sidewall according to the maximum ethanol accumulation rate limit. When the percentage of mitochondrial oscillation mode energy exceeds the threshold three times consecutively and the increase is greater than 17%, the baseline temperature of the ethanol tolerance thermal window is recorded.

6. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, The construction of the thermodynamic gradient for ethanol transmembrane transport includes: The theoretical upper limit of temperature difference can be calculated by combining the tank structure parameters with the heat flux density ratio; The temperature difference fluctuation range between the lower and upper temperature control zones is monitored. When the fluctuation range of 12 consecutive sampling points is less than ±0.07℃ and the average value is greater than 0.19℃, the temperature difference field is determined to have entered the initial steady state. The temperature of the lower temperature control zone is increased in a stepped manner while the temperature of the upper temperature control zone is adjusted simultaneously. When the five-stage temperature increase is completed and the energy ratio of the mitochondrial oscillation mode is stable, the width of the ethanol directional diffusion thermal window is determined.

7. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, Maintaining the directional diffusion thermal window of ethanol includes: The online metabolite monitoring unit was activated to continuously acquire transmission intensity and calculate the absorbance sequence and its second-order difference sequence. When the second-order difference sequence is positive for 5 consecutive times and the value is greater than 0.0023, it is determined that the formation of higher alcohol precursors has entered the accelerated phase. Adjust the heat flux density in the lower temperature control zone according to the acceleration level and heat flux sensitivity coefficient, while ensuring that the existing thermal window width is not damaged.

8. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, Maintaining mitochondrial membrane potential homeostasis includes: Calculate the tank's thermal inertia time constant by combining the parameters of the jacketed circulating medium; The standard deviation of mitochondrial membrane potential fluctuations was monitored. When three consecutive measurements were greater than 0.34 mV and increased by more than 22% compared to the previous value, the mitochondrial membrane stability was determined to be disturbed. Predict the rate of change of ethanol concentration in the future, issue a reverse heat flow compensation command in advance based on the thermal inertia time constant, and record the mitochondrial thermal steady-state reference temperature when the membrane potential fluctuation amplitude falls back to the stable range.

9. The dynamic temperature-controlled fermentation method according to claim 1, characterized in that, The controlled degradation of bioactivity includes: Using the mitochondrial thermal steady-state reference temperature as the target, the temperature difference for each region of the tank is calculated and the heat flow adjustment is allocated. Monitor the temperature deviation across the entire area, and determine that the temperature field homogenization is complete when the deviation is less than 0.09℃ for four consecutive times. The temperature is uniformly reduced to 12.4℃ at a rate of 0.13℃ / hour. When the temperature drops to 12.5℃, a final fermentation fingerprint package containing key parameters is generated, and the system switches to a cold standby mode.

10. A dynamic temperature-controlled fermentation system, characterized in that, Includes intelligent fermenters, online metabolite monitoring units, yeast activity sensing modules, and AI control platforms; The intelligent fermenter features an omnidirectional heat exchange wall and multi-point compatible interfaces; The online metabolite monitoring unit is configured to perform continuous transmission spectral acquisition; The yeast activity sensing module is configured to capture the extracellular adenosine triphosphate release flux and the amplitude of cell membrane potential fluctuations. The AI ​​control platform is configured to execute the dynamic temperature-controlled fermentation method according to any one of claims 1 to 9, generating heat flux density instructions based on the characteristics of wort and changes in yeast metabolic stages, so that temperature changes are coordinated and matched with the biological reaction process, achieving multi-stage, variable-speed, and adaptive temperature regulation.

Images