Liupu tea multi-modal micro-ecological directional fermentation bin and control method thereof

CN122997219APending Publication Date: 2026-10-30SOUTH ASIAN TROPICAL AGRI SCI RES INST OF GUANGXI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611172448.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-04
Publication Date
2026-10-30

AI Technical Summary

Technical Problem

即使个别方案能采集相关数据,也鲜少将风味接近度、微生态活性及异常风险同时纳入决策框架,且缺乏人工确认与PLC安全联锁优先级的分层控制机制

Benefits of technology

本发明通过多层分区料框与柔性散堆搅拌机构的协同设计,实现了发酵空间的灵活划分及散堆茶叶的均匀翻动与通气,显著提升了发酵环境的适应性与转化效率;结合复合感知机构获取的多维状态数据,工业控制柜能够基于FQI、MAI及RSI指数对发酵过程进行智能化的精准解析与定向调控,并通过人工确认与PLC安全联锁双重保障机制,有效规避了异常风险,实现了从环境参数到微生态活性的全流程数字化、精细化管理,从而确保了六堡茶发酵品质的稳定性与可控性,大幅提高了生产的自动化与安全水平。

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122997219A_ABST
    Figure CN122997219A_ABST
Patent Text Reader

Abstract

The application specifically relates to a Liupu tea multi-modal micro-ecological directional fermentation bin and a control method thereof. The fermentation bin comprises a heat preservation fermentation bin body, an internal multi-layer partitioned material frame, a partitioned environment adjusting mechanism, a composite sensing mechanism, a flexible bulk pile stirring mechanism and an industrial control cabinet. The material frame divides fermentation management zones, each of which is provided with an adjusting and sensing mechanism and is divided into a bulk pile fermentation management zone and a non-bulk pile fermentation management zone. The adjusting mechanism independently supplies air, discharges moisture, replenishes water or heats; the sensing mechanism collects temperature, humidity, gas, image and near-infrared data. The control cabinet is connected with each mechanism and executes a logic: according to data calculation, FQI (quality approaching index), MAI (micro-ecological activity index) and RSI (abnormal risk index) are calculated, a table is looked up to generate a suggestion, after manual confirmation and satisfaction of a PLC safety interlocking condition, driving execution is performed, and record feedback is provided. Through multi-layer partitioned independent control, multi-modal sensing feedback and intelligent index decision, the application realizes micro-ecological directional fermentation of Liupu tea.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the fields of tea fermentation equipment, food fermentation process control, and intelligent equipment for agricultural product processing, and more specifically to a multimodal microecological directional fermentation chamber for Liubao tea and its control method. Background Technology

[0002] Liubao tea is a post-fermented tea, and its core quality is highly dependent on the heat and moisture transformation, microbial metabolism, and changes in the contents of the tea leaves during the pile fermentation stage. During the fermentation process, factors such as pile temperature, moisture content, oxygen and carbon dioxide concentration, relative humidity, material thickness, turning time, water replenishment method, and ventilation directly determine the final product's aged aroma, areca nut aroma, mellowness, liquor color, and clarity.

[0003] Traditional Liubao tea fermentation relies primarily on experienced tea masters who judge the quality by "looking at the pile, touching the pile, smelling the aroma," and turning the pile. While this method offers flexibility for small-scale production, it exhibits significant instability in large-scale, standardized, and cross-seasonal production. Differences in personnel experience, uneven heat and humidity distribution in different locations within the storage area, and batch variations in the tea leaves can all lead to insufficient fermentation, over-fermentation, localized sourness, alcoholic odor, mold risk, or quality deviations.

[0004] Existing temperature and humidity control equipment typically only provides average control of the entire storage environment, lacking the ability to identify the local conditions of the tea leaves. However, significant spatial heterogeneity exists within the Liubao tea pile: temperature and humidity changes are inconsistent near the air ducts, doors, top, bottom, corners, and center. If control is based solely on single-point temperature and humidity or the average value of the entire storage, it is difficult to detect problems such as localized oxygen deficiency, excessive moisture, or excessively rapid temperature rise in a timely manner.

[0005] Furthermore, the existing fermentation equipment has a relatively coarse granularity in its water replenishment and ventilation. A single, full-warehouse water replenishment can easily lead to excessive surface moisture while the interior remains insufficiently moist; full-warehouse ventilation, on the other hand, may cause localized tea leaves to lose moisture too quickly. Liubao tea fermentation is better suited to differentiated processing based on zoned conditions, that is, adjusting air supply, dehumidification, and micro-mist water replenishment separately for different layers, columns, or material frames.

[0006] When the core temperature rises rapidly, traditional methods typically rely on manual turning or loosening of the pile. This approach is slow to respond, labor-intensive, and difficult to address localized overheating in deep piles or when space is limited. Using ordinary rigid stirring mechanisms can easily damage the tea leaves, increasing broken tea and affecting subsequent quality. Therefore, a flexible, loose-piling structure is urgently needed that can quickly loosen the material layer when overheated while minimizing mechanical damage to the tea leaves.

[0007] Meanwhile, given the lag in the formation of tea flavor, real-time temperature and humidity alone cannot directly determine whether quality parameters meet standards. Therefore, it is necessary to integrate multi-source information, including images, near-infrared spectroscopy, gas analysis, weight changes, sensor curves, human sensory evaluation, and quality inspection data, to establish an indicator system that reflects flavor quality, microecological activity, and abnormal risks. This will transform empirical judgments into process decisions that are visible, traceable, and reproducible.

[0008] Existing technologies have four main limitations: First, the equipment uses whole-warehouse temperature and humidity control, focusing only on the air environment inside the warehouse and unable to identify differences in the core of individual material frames or adjacent zones; second, the equipment relies on single-point or limited sensor monitoring, making it difficult to simultaneously acquire data from the surface and core of the stack, and thus unable to accurately determine local overheating, oxygen deficiency, or uneven moisture content; third, although the equipment includes turning or stirring mechanisms, these are mostly rigid turning or overall turning, making it difficult to balance rapid heat dissipation with maintaining the integrity of the tea leaves; fourth, although the digital systems involve data acquisition, microbial monitoring, or digital twin display, they usually remain at the level of displaying table entries or alarms, lacking a closed-loop control mechanism that directly converts data output into the actions of zoned air valves, nozzles, dehumidification, and flexible stacking mechanisms.

[0009] The aforementioned existing technologies generally lack a linkage mechanism for the three indicators FQI, MAI, and RSI that matches the fermentation process of Liubao tea. Even if some solutions can collect relevant data, they rarely incorporate flavor similarity, microecological activity, and abnormal risks into the decision-making framework simultaneously, and they lack a hierarchical control mechanism that prioritizes manual confirmation and PLC safety interlocking.

[0010] On the other hand, even if existing turning or stirring devices can change the state of the pile, they often do not take the integrity of tea leaves, low-speed torsion limitation, insertion depth limitation, and zoned wind and humidity linkage as unified design goals, making it difficult to simultaneously meet the dual requirements of rapid cooling and reducing breakage rate. Especially in pilot or large-scale production, without clear sensor placement rules, thresholds for judging differences between adjacent zones, and calibration mechanisms for table entries, the equipment is unlikely to achieve verifiable and replicable technical effects.

[0011] Therefore, there is an urgent need for a comprehensive technical solution that includes fermentation chamber structure, zoned execution mechanism, sensor deployment, three-index rule calculation, manual confirmation, and PLC safety interlock, in order to comprehensively improve the batch stability, risk warning capability, and quality reproducibility of Liubao tea fermentation. Summary of the Invention

[0012] The purpose of this invention is to provide a multimodal microecological directional fermentation chamber for Liubao tea and its control method. Through multi-layered independent control, multimodal perception feedback and intelligent indexed decision-making, the microecological directional fermentation of Liubao tea is realized.

[0013] To achieve the above objectives, the present invention provides the following technical solution: A multimodal microecological directional fermentation chamber for Liubao tea includes: a heat-insulated fermentation chamber body and a multi-layer partitioned material frame, a partitioned environmental adjustment mechanism, a composite sensing mechanism, a flexible random mixing mechanism, and an industrial control cabinet disposed in the heat-insulated fermentation chamber body. The multi-layer partitioned material frame is used to divide the fermentation space of the tea blank into multiple fermentation management zones. Each fermentation management zone is equipped with the partitioned environment adjustment mechanism and the composite sensing mechanism. The fermentation management zone equipped with the flexible loose-pile stirring mechanism is designated as the loose-pile fermentation management zone, while the fermentation management zone without the flexible loose-pile stirring mechanism is designated as the non-loose-pile fermentation management zone. The partitioned environmental control mechanism is used to independently regulate the air supply, dehumidification, water replenishment, or heating of each fermentation management zone. The composite sensing mechanism is used to collect fermentation status data of Liubao tea within the corresponding fermentation management area; wherein the fermentation status data includes: temperature, humidity, gas composition, image and near-infrared; The industrial control cabinet is electrically connected to the zoned environmental control mechanism, the composite sensing mechanism, and the flexible random mixing mechanism, respectively. The industrial control cabinet is configured to: calculate the FQI quality proximity index, MAI microecological activity index, and RSI abnormal risk index based on the sensing data; query the control suggestion mapping table to generate zonal control suggestions; after the zonal control suggestions are manually confirmed and meet the PLC safety interlock conditions, they are converted into execution outputs to drive the corresponding zonal environmental adjustment mechanism or flexible random mixing mechanism to perform actions, and the execution results and status feedback are recorded.

[0014] Furthermore, the zoned environmental control mechanism includes: a zoned air supply and dehumidification system and a zoned micro-mist water replenishment system; The zoned air supply and dehumidification system includes: a main air supply duct, zoned air supply branches corresponding to each fermentation management zone and zoned electric air valves, return air dehumidification channels and dehumidification outlets; The main air supply pipe is branched to multiple zone air supply branches, and each zone air supply branch corresponds to a fermentation management zone. The zoned electric air valves are installed on the air supply branch of each zone. The return air dehumidification channel is located on the return air side and connected to the dehumidification outlet; The zoned micro-mist water supply system includes: a micro-mist water supply main pipe and zoned micro-mist nozzles corresponding to each fermentation management zone; The main water supply pipe for micro-mist is connected to the corresponding micro-mist nozzles in each fermentation management area; The main water supply pipe for the micro-mist is equipped with a water supply valve or a metering pump to control the on / off supply or flow rate of water to the micro-mist nozzles in the designated areas.

[0015] Furthermore, the composite sensing mechanism includes: a composite sensor array, an image acquisition unit, a near-infrared acquisition unit, and a weighing base; The composite sensor array includes: a stack surface depth sensor and a stack core depth sensor; The surface depth sensor and the core depth sensor are used to collect data on temperature, relative humidity, tea leaf moisture, oxygen, carbon dioxide, volatile organic compounds and pressure difference of the surface and core of the fermented tea pile, respectively. The image acquisition device is used to collect the surface color, material layer flatness, and clumping state of the tea leaves; The near-infrared collector is used to collect the near-infrared reflection characteristics and near-infrared principal component scores of the tea blank. The weighing base is used to collect the weight data of the tea leaves; The surface depth sensor is set at a position 20-50mm from the surface of the tea blank, and the core depth sensor is set at a position 150-300mm from the surface of the tea blank or at 1 / 2-2 / 3 of the thickness of the material layer; each layer of material frame is equipped with no less than 4 sets of sensor points.

[0016] Furthermore, the flexible random mixing mechanism includes: a deceleration drive assembly, a low-speed drive shaft, a flexible rounded corner paddle, a torque limiting protection component, and a position limiting component; The speed reduction drive assembly includes a drive motor and a speed reducer, which are located outside the partitioned material frame in the dispersible fermentation management area and fixed to the side wall of the partitioned material frame or to the mounting bracket connected to the partitioned material frame. The output end of the drive motor is connected to the input end of the reducer, and the output end of the reducer is connected to one end of the low-speed drive shaft through a torque limiting protection component. The low-speed drive shaft is arranged laterally along the partitioned material frame and extends into or above the tea blank layer. The multiple flexible rounded corner paddles are fixedly connected to the shaft segment of the low-speed drive shaft located within the partitioned material frame, and are arranged axially and circumferentially staggered along the low-speed drive shaft, so that when the low-speed drive shaft rotates, it drives the flexible rounded corner paddles to flip the tea blanks. The position limiting component is disposed on the side wall of the partitioned material frame or on the mounting bracket of the low-speed drive shaft, and is limited and cooperated with the mounting part of the low-speed drive shaft. The low-speed drive shaft has a rotational speed range of 2-30 rpm, the torque limiting protection component has a protection torque range of 0.5-8 N·m, and the position limiting component is used to limit the insertion depth of the flexible rounded corner paddle to no more than two-thirds of the material layer thickness.

[0017] Furthermore, the industrial control cabinet includes: an industrial computing unit, a PLC safety control module, a process parameter and table entry storage module, and a communication gateway and data traceability module; The industrial computing unit is used to calculate the FQI quality proximity index, MAI microecological activity index and RSI abnormal risk index based on the sensing data, and to query the control suggestion mapping table to generate zonal control suggestions. The process parameter and table storage module is used to store the target process-quality parameter table, the index weight table, the control suggestion mapping table, and the PLC safety threshold. The PLC safety control module is used to receive manual confirmation signals and perform safety interlock judgments. When the manual confirmation signal is valid and the safety interlock conditions are met, the zoning control suggestion is converted into an execution output. The PLC safety interlock has a higher priority than the zoning control suggestion output by the industrial computing unit. When the process parameter to be judged is within the safe range but exceeds the allowable range, the zoning control suggestion is not converted into an execution output, but only an alarm prompt is generated. When the process parameter to be judged exceeds the safe range, the corresponding automatic execution action is prohibited. The communication gateway and data traceability module are used to record batch data, recommended actions, manual confirmation records, PLC execution results, and execution status feedback.

[0018] Furthermore, a waterproof junction box and cable tray assembly for the insulated fermentation chamber is provided between the outer wall of the chamber and the industrial control cabinet. The waterproof junction box and cable tray assembly for the fermentation chamber-control cabinet includes: an inlet terminal located on the outer wall of the insulated fermentation chamber and on the side of the control cabinet; The signal lines, drive lines, feedback lines, and safety interlock lines of the composite sensing mechanism, the zoned environmental control mechanism, and the flexible random mixing mechanism are all connected to the industrial control cabinet via the waterproof junction box and cable tray assembly of the silo-control cabinet.

[0019] This invention also provides a control method for a multimodal microecological directional fermentation chamber for Liubao tea, the method comprising: S1. Real-time collection of fermentation status data of Liubao tea in the corresponding fermentation management area through a composite sensing mechanism; the fermentation status data includes: temperature, humidity, gas composition, image and near-infrared data; S2. Calculate the FQI quality proximity index, MAI microecological activity index, and RSI abnormal risk index based on the perceived data. S3. Query the control suggestion mapping table and generate partition control suggestions based on the calculation results; S4. Manually confirm the proposed partition control and determine whether the PLC safety interlock conditions are met. S5. If manual confirmation is passed and the PLC safety interlock conditions are met, the zoning control suggestion will be converted into an execution output to drive the corresponding zoning environmental regulation mechanism to perform independent air supply, dehumidification, water replenishment or heating regulation actions, or drive the flexible bulk mixing mechanism to perform mixing actions in the bulk fermentation management area. S6. Record execution results and status feedback.

[0020] Furthermore, the formula for calculating FQI is as follows: FQI=100*(w1C+w2A+w3T+w4M+w5N+w6Q); In the formula, C represents color similarity, A represents VOC aroma vector similarity, T represents near-infrared taste calibration similarity, M represents tea leaf moisture suitability, N represents near-infrared maturity similarity, and Q represents historical quality inspection matching degree; C, T, M, and N are all target interval type suitability parameters, A is converted according to the weighted distance between the current VOC vector and the target VOC vector, and Q is converted according to the weighted distance between the current batch feature vector and the historical qualified batch mean vector; w1, w2, w3, w4, w5, and w6 are weight coefficients and their sum is 1.

[0021] Furthermore, the formula for calculating MAI is as follows: MAI=100*(v1R+v2W+v3O+v4G+v5V+v6B); Wherein, R is the suitability of the core heating rate, W is the suitability of the water content change per unit time, O is the oxygen consumption characteristic per unit time, G is the carbon dioxide generation characteristic per unit time, V is the VOC change characteristic per unit time, and B is the microecological verification characteristic; each component is normalized to 0-1; when no verification data is available, the weight of item B is proportionally allocated to other items.

[0022] Furthermore, the formula for calculating the RSI is as follows: RSI=100*max(E,H,L r G c V r ,S,D); Where E represents overheating risk, H represents overhumidity risk, and L represents overheating risk. r For the risk of hypoxia, Gc is for the risk of CO2 accumulation, and V is for the risk of oxygen deficiency. r S represents abnormal VOC risk, D represents sensor failure risk, and D represents actuator abnormal risk. Each individual risk value is converted to 0-1 based on the degree to which the parameter deviates from the target range or safe range.

[0023] According to specific embodiments provided by the present invention, the present invention has the following technical effects compared to the prior art: This invention, through the collaborative design of a multi-layered partitioned material frame and a flexible loose-pile stirring mechanism, achieves flexible division of the fermentation space and uniform turning and aeration of loose-pile tea leaves, significantly improving the adaptability and conversion efficiency of the fermentation environment. Combined with multi-dimensional state data acquired by a composite sensing mechanism, the industrial control cabinet can perform intelligent and precise analysis and directional control of the fermentation process based on FQI, MAI, and RSI indices. Through a dual guarantee mechanism of manual confirmation and PLC safety interlock, abnormal risks are effectively avoided, realizing full-process digital and refined management from environmental parameters to microecological activity. This ensures the stability and controllability of Liubao tea fermentation quality and greatly improves the automation and safety level of production. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0025] The following description, in conjunction with the accompanying drawings, further illustrates the present invention of a multimodal microecological directional fermentation chamber for Liubao tea and its control method. Figure 1 This is a schematic diagram of the overall structure of the fermentation chamber in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the appearance and external connections of the industrial control cabinet in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of the internal electrical module arrangement of the industrial control cabinet in Embodiment 1 of the present invention; Figure 4 This is a side view sectional disassembly diagram of the fermentation chamber in Embodiment 1 of the present invention; Figure 5 This is a partial schematic diagram of the dual-depth sensor layout in Embodiment 1 of the present invention; Figure 6 This is a partial schematic diagram of the flexible random mixing mechanism in Embodiment 1 of the present invention; Figure 7 This is a schematic diagram of a single-layer, four-zone flexible random stacking arrangement in Embodiment 1 of the present invention; Figure 8 This is a schematic diagram of the control closed-loop signal relationship in Embodiment 1 of the present invention; Figure 9 This is a schematic diagram of the connection relationship of the flexible random mixing mechanism in Embodiment 1 of the present invention.

[0026] In the attached diagram: 1. Insulated fermentation chamber; 2. Sealed chamber door; 3. Multi-layer material cart; 4. Zoned material frame; 5. Main air supply duct; 6. Zoned air supply branch; 7. Zoned electric air valve; 8. Return air dehumidification channel; 9. Dehumidification outlet; 10. Micro-mist water supply main duct; 11. Zoned micro-mist nozzle; 12. Composite sensor array; 121. Surface depth sensor; 122. Core depth sensor; 123. Zoned sensor bracket; 124. Material layer thickness direction indicator; 125. Surface sensor installation depth scale; 126. Core sensor installation depth scale; 13. Image / near-infrared acquisition device; 14. 15. Weighing base; 16. Heating and dehumidification module; 17. Industrial control cabinet; 18. Sensor input terminal block; 19. Image / near-infrared acquisition interface; 10. Microbial / quality inspection data interface; 11. Industrial computing unit; 12. Process parameter and table storage module; 13. FQI / MAI / RSI calculation and display module; 14. Batch and fermentation management area selection display area; 15. FQI / MAI / RSI index display area; 16. FQI / MAI / RSI suggestion display area; 16. 5. PLC Safety Interlock and Alarm Status Display Area; 1607. PLC Safety Control Module; 1608. Safety Relay and Emergency Stop Circuit; 1609. Zoned Air Valve Drive Output Module; 1610. Nozzle / Metering Pump Drive Output Module; 1611. Agitator Driver / Inverter; 1612. Execution Status Feedback Terminal Block; 1613. Manual Confirmation and Alarm Reset Unit; 1614. Communication Gateway and Data Traceability Module; 1615. Power Isolation and Grounding Module; 17. PLC Safety Interlock Status / Interface Area; 18. Communication Gateway Interface Area; 19. Data Traceability Interface Area; 2 0. Cleaning drain outlet; 21. Emergency stop button; 22. Audible and visual alarm; 23. Flexible bulk mixing mechanism; 231. Low-speed drive shaft; 232. Flexible rounded corner paddle; 233. Torque limiting protection component; 234. Position limiting component; 235. Detachable cleaning interface; 236. Speed ​​detection component; 237. Torque detection component; 238. Insertion depth limiting component; 239. Reduction drive assembly; 24. Waterproof junction box and cable tray assembly for silo-control cabinet; 240. Bearing support seat; 241. Food contact grade sealing ring; 242. Bulk fermentation management area; 243. Non-bulk fermentation management area. Detailed Implementation

[0027] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0028] To better understand the purpose, structure, and function of this invention, the invention will be described in further detail below with reference to the accompanying drawings.

[0029] Example 1 like Figure 1 As shown, the present invention provides a multimodal micro-ecological directional fermentation chamber for Liubao tea, including: a heat-insulated fermentation chamber 1 and a multi-layer partitioned material frame 4, a partitioned environment adjustment mechanism, a composite sensing mechanism, a flexible random pile stirring mechanism, and an industrial control cabinet disposed within the heat-insulated fermentation chamber 1. The multi-layer partitioned material frame 4 is used to divide the fermentation space of Liubao tea into multiple fermentation management zones. Each fermentation management zone is equipped with the partitioned environment adjustment mechanism and the composite sensing mechanism. The fermentation management zone equipped with the flexible loose-pile stirring mechanism 23 is designated as the loose-pile fermentation management zone 242, and the fermentation management zone without the flexible loose-pile stirring mechanism is designated as the non-loose-pile fermentation management zone 243. In this embodiment, the multi-layer material cart 3 and the partitioned material frame 4 are arranged in the heat-insulated fermentation chamber 1. The multi-layer material cart 3 is a frame structure that can enter and exit the heat-insulated fermentation chamber 1. The partitioned material frame 4 is arranged in layers and columns to form no less than 8 fermentation management areas. Each fermentation management area has a detachable material frame, a breathable support bottom surface, a sensor mounting position, a nozzle mounting position, and a drainage guide.

[0030] The flexible bulk mixing mechanism 23 includes: a deceleration drive assembly 239, a low-speed drive shaft 231, a flexible rounded corner paddle 232, a torque limiting protection component 233, and a position limiting component 234; In this embodiment, the drive motor and the reducer constitute a speed reduction drive assembly 239, which is disposed outside the partitioned material frame in the dispersible fermentation management area and fixed to the side wall of the partitioned material frame or to a mounting bracket connected to the partitioned material frame. The output end of the drive motor is drivenly connected to the input end of the reducer, and the output end of the reducer is drivenly connected to one end of the low-speed drive shaft 231 through a torque limiting protection component 233, so as to limit the torque transmitted to the low-speed drive shaft 231 when stall occurs or the resistance is too high.

[0031] The low-speed drive shaft 231 is arranged laterally along the partitioned material frame and extends into or above the tea blank material layer. Multiple flexible rounded corner paddles 232 are fixedly connected to the shaft section of the low-speed drive shaft 231 located in the partitioned material frame, and are arranged axially spaced and circumferentially staggered along the low-speed drive shaft 231, so that when the low-speed drive shaft 231 rotates, it drives the flexible rounded corner paddles 232 to push, lift, or turn the tea blank.

[0032] Position limiting component 234 is disposed on the side wall of the partitioned material frame or on the mounting bracket of the low-speed drive shaft 231, and cooperates with the mounting part of the low-speed drive shaft 231 to limit the position of the low-speed drive shaft 231 and the flexible rounded corner paddle 232 relative to the tea material layer, so that the insertion depth of the flexible rounded corner paddle 232 into the tea material layer does not exceed two-thirds of the material layer thickness. The drive motor is also electrically connected to the stirring driver / frequency converter in the industrial control cabinet through drive circuit.

[0033] At least one fermentation management area is equipped with the flexible bulk mixing mechanism 23 and serves as a bulk fermentation management area 242; other fermentation management areas without the flexible bulk mixing mechanism are designated as non-bulk fermentation management areas 243.

[0034] The low-speed drive shaft 231 has a rotational speed range of 2-30 rpm, the torque limiting protection component 233 has a protection torque range of 0.5-8 N·m, and the position limiting component 234 is used to limit the insertion depth of the flexible rounded corner paddle 232 to no more than two-thirds of the material layer thickness.

[0035] The flexible rounded corner paddle 232 is made of food-contact grade elastic material or a rounded corner paddle covered with a flexible layer. The material hardness is Shore A40-80, the radius of the rounded corner of the paddle edge is not less than 2mm, the installation angle between the paddle and the low-speed drive shaft 231 is 20-60 degrees, and multiple paddles are arranged alternately along the low-speed drive shaft 231 and act on the tea blank by pushing, lifting or flipping.

[0036] The partitioned environmental control mechanism is used to independently regulate the air supply, dehumidification, water replenishment, or heating of each fermentation management zone. The zoned environmental control mechanism includes: a zoned air supply and dehumidification system and a zoned micro-mist water replenishment system; The zoned air supply and dehumidification system includes: a main air supply pipe 5, zoned air supply branches 6 corresponding to each fermentation management zone, zoned electric air valves 7, return air dehumidification channels 8, and dehumidification outlets 9. The main air supply pipe 5 is branched to multiple zone air supply branches 6, and each zone air supply branch 6 corresponds to a fermentation management zone. The partition electric air valve is installed on the air supply branch of each fermentation management area; each partition electric air valve 7 corresponding to each fermentation management area has an opening feedback unit, which is connected to the execution status feedback terminal block. The return air dehumidification channel 8 is located on the return air side and connected to the dehumidification port 9; The zoned micro-mist water supply system is located at the top or side of each fermentation management zone and includes: a micro-mist water supply main pipe 10 and zoned micro-mist nozzles 11 corresponding to each fermentation management zone; The main water supply pipe 10 for micro-mist is connected to the corresponding micro-mist nozzles 11 in each fermentation management area; The main water supply pipe 10 for micro-mist is equipped with a water supply valve or a metering pump to control the on / off state or flow rate of water supply to the micro-mist nozzles in the zones.

[0037] The composite sensing mechanism is used to collect fermentation status data of Liubao tea within the corresponding fermentation management area; wherein the fermentation status data includes: temperature, humidity, gas composition, image and near-infrared; The composite sensing mechanism includes: a composite sensor array, an image acquisition unit, a near-infrared acquisition unit, and a weighing base; The composite sensor array includes: a surface depth sensor 121 and a core depth sensor 122; The surface depth sensor 121 and the core depth sensor 122 are used to collect data on temperature, relative humidity, tea leaf moisture, oxygen, carbon dioxide, volatile organic compounds and pressure difference of the surface and core of the fermented tea pile, respectively. The image acquisition device is used to collect the surface color, material layer flatness, and clumping state of the tea leaves; The near-infrared collector is used to collect the near-infrared reflection characteristics and near-infrared principal component scores of the tea blank. In this embodiment, the image / near-infrared collector 13 is used to collect the surface color of the tea leaves, the flatness of the material layer, the clumping state, the near-infrared reflection characteristics, and the near-infrared principal component score; the microecological / quality inspection data interface 1603 is used to access at least one of the following verification data: ATP, pH, conductivity, colony count, dominant microbial community detection, or physicochemical quality inspection results.

[0038] The weighing base 14 is used to collect the weight data of the tea blanks; The surface depth sensor 121 is located 20-50mm from the surface of the tea blank, and the core depth sensor 122 is located 150-300mm from the surface of the tea blank or at 1 / 2-2 / 3 of the thickness of the material layer; each layer of the material frame is equipped with no less than 4 sets of sensor points.

[0039] The industrial control cabinet 16 is electrically connected to the zoned environmental control mechanism, the composite sensing mechanism, and the flexible random mixing mechanism 23, respectively. The industrial control cabinet 16 is configured to execute the following control logic: calculate the FQI quality proximity index, MAI microecological activity index and RSI abnormal risk index based on the sensing data, query the control suggestion mapping table to generate zonal control suggestions; after the zonal control suggestions are manually confirmed and meet the PLC safety interlock conditions, they are converted into the execution output to drive the corresponding zonal environmental adjustment mechanism or flexible random mixing mechanism 23 to act, and the execution results and status feedback are recorded.

[0040] The control recommendation mapping table configures the target actuator field according to the fermentation management area. The target actuator field of the free-floating fermentation management area 242 includes a stirring drive / frequency converter, and the target actuator field of the non-free-floating fermentation management area 243 includes at least one of the following: zone air valve, return air dehumidification channel, water supply valve or metering pump. Figure 7A1 and A3 are examples of reclaimed fermentation management areas with flexible reclaimed mixing mechanisms installed, while A2 and A4 are examples of non-reclaimed fermentation management areas without flexible reclaimed mixing mechanisms installed; A1-A4 are example partition numbers for a single-layer four-partition system.

[0041] The industrial control cabinet 16 includes: an industrial computing unit 1604, a PLC safety control module 1607, a process parameter and table entry storage module 1605, and a communication gateway and data traceability module 1614. The industrial computing unit 1604 is used to calculate the FQI quality proximity index, MAI microecological activity index and RSI abnormal risk index based on the sensing data, and to query the control suggestion mapping table to generate zoning control suggestions. The process parameter and table storage module 1605 is used to store the target process-quality parameter table, the index weight table, the control suggestion mapping table and the PLC safety threshold. The PLC safety control module 1607 is used to receive manual confirmation signals and perform safety interlock judgments. When the manual confirmation signal is valid and the safety interlock conditions are met, the zoning control suggestion is converted into an execution output. The PLC safety interlock has a higher priority than the zoning control suggestion output by the industrial computing unit 1604. When the process parameter to be judged is within the safe range but exceeds the allowable range, the zoning control suggestion is not converted into an execution output, but only an alarm prompt is generated. When the process parameter to be judged exceeds the safe range, the corresponding automatic execution action is prohibited. The communication gateway and data traceability module 1614 are used to record batch data, recommended actions, manual confirmation records, PLC execution results, and execution status feedback.

[0042] A waterproof junction box and cable tray assembly 24 is provided between the outer wall of the insulated fermentation chamber 1 and the industrial control cabinet 16. The waterproof junction box and cable tray assembly 24 for the silo-control cabinet includes: a waterproof junction box on the outer wall of the silo, an inlet terminal on the control cabinet side, and a cable tray, conduit or cable bundle connecting the two.

[0043] In this embodiment, the waterproof junction box and cable tray assembly 24 of the silo-control cabinet is disposed between the outer wall of the insulated fermentation silo 1 and the industrial control cabinet 16, and includes at least one of the following: a waterproof junction box on the silo side, a cable inlet on the control cabinet side, and a cable tray, conduit or cable bundle connected between the two. The industrial control cabinet 16 includes: a sensor input terminal block 1601, an image / near-infrared acquisition interface 1602, a micro-ecological / quality inspection data interface 1603, an industrial computing unit 1604, a process parameter and table entry storage module 1605, an FQI / MAI / RSI calculation and display module 1606, a PLC safety control module 1607, a safety relay and emergency stop circuit 1608, a zoned air valve drive output module 1609, a nozzle / metering pump drive output module 1610, a stirring driver / frequency converter 1611, an execution status feedback terminal block 1612, a manual confirmation and alarm reset unit 1613, a communication gateway and data traceability module 1614, and a power isolation and grounding module 1615; The process parameter and table storage module 1605 stores the target process-quality parameter table, the index weight table, the control suggestion mapping table, and the PLC safety threshold; the industrial computing unit 1604 calculates the FQI quality proximity index, MAI microecological activity index, and RSI abnormal risk index based on the data input from the composite sensor array 12, the image / near-infrared collector 13, the weighing base 14, and the microecological / quality inspection data interface 1603, and generates a partitioned control suggestion containing the target actuator field and the execution parameter field based on the control suggestion mapping table; The zoning control suggestions are first displayed on the FQI / MAI / RSI calculation display module 1606 and do not directly drive the actuators. The PLC safety control module 1607 only converts the zoning control suggestions into the execution outputs of the corresponding fermentation management area's zoning electric air valve 7, zoning micro-mist nozzle 11, heating and dehumidification module 15, or flexible random stirring mechanism 23 when the manual confirmation signal is valid and the PLC safety interlock meets the conditions. The outputs are then fed back to the corresponding actuators via the execution status feedback terminal block 1612, showing the opening / closing status, opening degree, flow rate, speed, torque, or alarm status. The communication gateway and data traceability module 1614 records batch data, recommended actions, manual confirmation records, PLC execution results, and execution status feedback. The power isolation and grounding module 1615 connects the safety relay and emergency stop circuit 1608 and is used to provide power isolation, grounding, or safe power supply protection.

[0044] The FQI / MAI / RSI calculation display module 1606 includes a batch and fermentation management area selection display area 16061 visible from the cabinet door, a zone real-time parameter display area 16062, an FQI / MAI / RSI index display area 16063, a zone control suggestion display area 16064, and a PLC safety interlock and alarm status display area 16065. The outside of the industrial control cabinet door is also equipped with a PLC safety interlock status / interface area 17, a communication gateway interface area 18, and a data traceability interface area 19. The PLC safety interlock status / interface area 17 is connected to the PLC safety control module 1607 and the safety relay and emergency stop circuit 1608. The communication gateway interface area 18 and the data traceability interface area 19 are connected to the communication gateway and data traceability module 1614. The control suggestion mapping table includes: an input condition field, a risk level field, an execution output field, a PLC allowable condition field, and a prohibitive execution condition field. The execution output field includes a target actuator field and an execution parameter field.

[0045] The PLC safety control module 1607 has interlock protection logic for emergency stop, over-temperature, over-humidity, oxygen deficiency, electrical abnormality, air valve abnormality, nozzle abnormality, drive shaft overload, door not closed, and communication abnormality. The PLC safety interlock has a higher priority than the zoning control suggestions output by the industrial computing unit.

[0046] Example 2 The present invention also provides a control method for the multimodal microecological directional fermentation chamber for Liubao tea in Example 1, the method comprising: S1. Real-time collection of fermentation status data of Liubao tea in the corresponding fermentation management area through a composite sensing mechanism; the fermentation status data includes: temperature, humidity, gas composition, image and near-infrared data; S2. Calculate the FQI quality proximity index, MAI microecological activity index, and RSI abnormal risk index based on the perceived data. S3. Query the control suggestion mapping table and generate partition control suggestions based on the calculation results; S4. Manually confirm the proposed partition control and determine whether the PLC safety interlock conditions are met. S5. If manual confirmation is passed and the PLC safety interlock conditions are met, the zoning control suggestion is converted into an execution output, driving the corresponding zoning environmental adjustment mechanism to perform independent air supply, dehumidification, water replenishment or heating adjustment actions, or driving the flexible bulk mixing mechanism 23 to perform mixing actions in the bulk fermentation management area. S6. Record execution results and status feedback.

[0047] In this embodiment, the method specifically involves: establishing a batch file containing tea batch, loading amount, target process-quality parameter table, index weight table, control suggestion mapping table, and PLC safety threshold; collecting temperature, relative humidity, tea moisture, oxygen, carbon dioxide, VOC, pressure difference, weight, image, and near-infrared data at the surface depth and core depth of the fermentation management area; selecting numerical parameters that can be written into the target process-quality parameter table and have target range, allowable range, and safe range as process parameters to be judged, including temperature, relative humidity, tea moisture, oxygen, carbon dioxide, VOC characteristic values, pressure difference, weight, color similarity or clumping characteristic values ​​converted from image data, maturity similarity or principal component score converted from near-infrared data, and the difference values ​​between adjacent fermentation management areas formed by the above numerical parameters; and performing anomaly removal, amplitude normalization, and adjacent zone difference analysis on the process parameters to be judged based on the target process-quality parameter table and index weight table. The system verifies and calculates the FQI (Quality Approach Index), MAI (Microecological Activity Index), and RSI (Abnormal Risk Index). Based on the FQI, MAI, RSI, the process parameters to be judged, and the differences between adjacent fermentation management areas, it queries the control suggestion mapping table to generate corresponding zoning control suggestions for the fermentation management area. The zoning control suggestions include target actuator fields and execution parameter fields and are initially displayed for manual confirmation. When the process parameters to be judged are within the target range or allowable range, the manual confirmation signal is valid, and the PLC safety interlock meets the conditions, the PLC safety control module 1607 executes the corresponding action and records the execution result and feedback signal. When the process parameters to be judged are within the safe range but exceed the allowable range, the PLC safety control module 1607 does not convert the zoning control suggestion into an execution output, but only generates an alarm prompt and a manual review request. When the process parameters to be judged are less than the lower safety limit SL or greater than the upper safety limit SU, or when the PLC safety threshold is triggered, the PLC safety control module prohibits the corresponding automatic execution action.

[0048] The process parameters to be judged refer to numerical parameters that can be written into the target process-quality parameter table and have target range, allowable range and safe range. These parameters include at least one of the following: temperature, relative humidity, tea leaf moisture, oxygen, carbon dioxide, VOC characteristic value, pressure difference, weight, color similarity or clumping characteristic value converted from image data, maturity similarity or principal component score converted from near-infrared data, and difference values ​​between adjacent fermentation management zones formed by the above parameters.

[0049] The target process-quality parameter table includes: parameter name, target lower limit L, target upper limit U, allowable lower limit AL, allowable upper limit AU, safety lower limit SL and safety upper limit SU; wherein SL is less than or equal to AL, AL is less than L, L is less than or equal to U, U is less than AU, and AU is less than or equal to SU.

[0050] The target process-quality parameter table stores parameter names, target ranges, allowable ranges, safe ranges, maximum allowable VOC distance, maximum allowable distances for historical quality inspection features, and target ranges for near-infrared features or principal component scores. The index weight table stores the component weights of FQI, MAI, and RSI, as well as the order of each vector component. The control suggestion mapping table stores input conditions, risk levels, execution output fields, PLC allowable conditions, and prohibitive execution conditions. The execution output fields include target actuator fields and execution parameter fields. These three tables, along with the PLC safety thresholds, are stored together in the process parameter and table item storage module.

[0051] For the target interval type suitability parameter, the score is 1 when x is in [L,U], the score is calculated as (x-AL) / (L-AL) when x is in [AL,L), the score is calculated as (AU-x) / (AU-U) when x is in (U,AU), the score is calculated as (AU-x) / (AU-U) when x is in [SL,AL) or (AU,SU], the score is 0 and a manual confirmation prompt is triggered, and the PLC prohibition condition is triggered when x is less than SL or greater than SU.

[0052] The formula for calculating FQI is as follows: FQI=100*(w1C+w2A+w3T+w4M+w5N+w6Q) In the formula, C represents color similarity, A represents VOC aroma vector similarity, T represents near-infrared taste calibration similarity, M represents tea leaf moisture suitability, N represents near-infrared maturity similarity, and Q represents historical quality inspection matching degree; A is calculated based on the current VOC vector X and the target VOC vector T. g The weighted distance dA is converted, dA=s qrt (sum(k a *(X a -T ga )^2)), A=1-min(d A / dA max ,1), where k a Here, dA represents the weight of the VOC component in the indicator weight table. max The maximum allowable distance for VOCs and dA recorded in the target process-quality parameter table. max >0; when dA max When =0, A does not enter automatic judgment and triggers manual review; Q compares the feature vector Y composed of the current batch's moisture, color, VOC, and near-infrared characteristics with the average vector Y of similar qualified batches. g Weighted distance dQ conversion; dQ=s qrt (sum(k q *(Y q -Y gq)^2)), Q=1-min(dQ / dQ max ,1); Where k q dQ represents the weights of the historical quality inspection features in the indicator weight table. max The maximum allowable distance and dQ for the same type of qualified batches recorded in the target process-quality parameter table. max >0; when dQ max When =0, Q is calculated using a normalized weighted average or equal weighted average of C, A, T, M, and N; C, T, M, and N are target interval suitability parameters converted to 0-1 according to the scoring rules; where C is obtained by comparing the L*a*b* value, RGB value, or color difference value after whiteboard correction with the target color interval, T is obtained by comparing the near-infrared characteristic band or principal component score with the target taste calibration interval, M is obtained by comparing the moisture content of the tea base with the target moisture interval, and N is obtained by comparing the near-infrared maturity characteristic value with the target maturity interval; where VOC vector X and target VOC vector T g Including the same number and in the same order of VOC components, X a and T ga Let sum(k) be the current value and the target value of the a-th VOC component, respectively. a *(X a -T ga )^2) is the summation of each VOC component a; the feature vector Y and the mean vector Yg of the same type of qualified batches include the same number and consistent order of historical quality inspection features, Y q and Y gq Let q be the current value and mean of the q-th historical quality inspection feature, respectively. sum(kq*(Yq-Ygq)^2) is the summation over each historical quality inspection feature q. C, A, T, M, N, and Q are all normalized to 0-1. w1-w6 are weighting coefficients greater than or equal to 0, and w1+w2+w3+w4+w5+w6=1. When there are no historical samples of similar qualified batches, Q is replaced by the normalized weighted average of C, A, T, M, and N. Q=(w1C+w2A+w3T+w4M+w5N) / (w1+w2+w3+w4+w5; Where w1+w2+w3+w4+w5>0; when w1+w2+w3+w4+w5=0, Q is calculated with equal weights for the five items C, A, T, M and N, and the status of no historical samples is recorded in the batch file.

[0053] The MAI is calculated as MAI=100*(v1R+v2W+v3O+v4G+v5V+v6B); In the formula, R represents the suitability of the core heating rate, W represents the suitability of the water content change per unit time, O represents the oxygen consumption characteristic per unit time, G represents the carbon dioxide generation characteristic per unit time, V represents the VOC change characteristic per unit time, and B represents at least one verification characteristic among ATP, pH, conductivity, colony count, dominant microbial community detection, or physicochemical quality inspection results; R, W, O, G, and V are converted to 0-1 according to the target range, allowable range, or unit time change threshold in the target process-quality parameter table, and B is converted to 0-1 according to the accessed verification data relative to the corresponding verification threshold or qualified range; v1-v6 are the original weight coefficients greater than or equal to 0 and v1+v2+v3+v4+v5+v6=1.

[0054] When the verification data corresponding to B is not connected, the real-time calculation weight of item B is reset to 0, and the original weight v6 of item B in the indicator weight table is allocated according to the original weight ratios of R, W, O, G, and V, where v1+v2+v3+v4+v5>0: v1'=v1+v6*v1 / (v1+v2+v3+v4+v5); v2'=v2+v6*v2 / (v1+v2+v3+v4+v5); v3'=v3+v6*v3 / (v1+v2+v3+v4+v5); v4'=v4+v6*v4 / (v1+v2+v3+v4+v5); v5'=v5+v6*v5 / (v1+v2+v3+v4+v5); The redistributed weights satisfy: v1'+v2'+v3'+v4'+v5'=1; And calculate according to MAI=100*(v1'R+v2'W+v3'O+v4'G+v5'V); When v1+v2+v3+v4+v5=0, MAI is calculated with equal weights for the five terms R, W, O, G and V.

[0055] The RSI is defined as RSI=100*max(E,H,L) r G c V r Calculate (S,D); In the formula, E represents overheating risk, H represents overhumidity risk, Lr represents hypoxia risk, and G represents... c Accumulating risks for CO2, V r S represents abnormal VOC risk, D represents sensor failure risk, and E represents the maximum value among core temperature risk, surface-to-core temperature difference risk, and heating rate risk. H represents the maximum value among relative humidity risk, tea leaf moisture risk, and water replenishment feedback risk. rG is the maximum of the risk of O2 concentration and the risk of O2 decline rate. c V is the maximum value between the risk of CO2 concentration and the risk of CO2 rise rate. r Take the maximum value among the deviation risks of each VOC feature; the above individual risks are divided into high-value risk type, low-value risk type and interval deviation type according to the direction of parameter risk and converted into 0-1 risk value; For high-risk parameters, the risk value is 0 when x is below the target upper limit U, calculated as (xU) / (AU-U) when x is within (U,AU), and 1 when x is within (AU,SU), triggering a manual confirmation prompt. If x is greater than SU, a PLC execution prohibition condition is triggered. For low-risk parameters, the risk value is 0 when x is above the target lower limit L, calculated as (Lx) / (L-AL) when x is within [AL,L), and 1 when x is within [SL,AL), triggering a manual confirmation prompt. If x is less than SL, a PLC execution prohibition condition is triggered. For interval deviation parameters, the risk value is 0 when the parameter is within the target interval [L,U], and is linearly calculated based on the distance from the corresponding target boundary when it is within [AL,L) or (U,AU]. The risk value is 1 when the parameter is within [SL,AL) or (AU,SU] and triggers a manual confirmation prompt. When the value is less than SL or greater than SU, the PLC execution prohibition condition is triggered. S is normalized to 0-1 based on the number of continuous sampling cycles for consecutive missing, out-of-bounds, or inconsistent adjacent depth readings. D is normalized to 0-1 based on the alarm status of abnormal air valves, nozzles, metering pumps, stirring drives, or feedback terminals.

[0056] FQI is used to characterize the degree of proximity to quality targets, MAI is used to characterize microecological metabolic activity, and RSI is used to characterize the level of abnormal risk. FQI, MAI, and RSI are calculated by the industrial computing unit according to the target process-quality parameter table, the indicator weight table, and the control recommendation mapping table. Their components are all derived from sensor data, image / near-infrared data, VOC data, weighing data, and optional quality inspection and verification data. The VOC vector, historical quality inspection feature vector, near-infrared feature, the order of each component, the weight, and the maximum allowable distance are all pre-written into the corresponding table entries or written by the quality inspection and calibration process after the batch ends. Undefined abstract evaluations do not replace specific data fields.

[0057] When the moisture content in a certain fermentation management area is below the target lower limit and the moisture difference between adjacent areas reaches 2-4 percentage points, the target actuator is the water supply valve or metering pump for that area; when the CO2 concentration difference reaches 1000-3000 ppm or the O2 concentration difference reaches 1-3 percentage points, the target actuator is the air valve and return air dehumidification channel for that area; when the core temperature is 2-7℃ higher than the target core temperature upper limit, the temperature difference between the stack surface and the core reaches 5-12℃, or the heating rate reaches 1.5-3.0℃ per hour, the target actuator is the flexible stack mixing mechanism of the dispersible fermentation management area, which is linked to the zone's air supply and dehumidification. When the core temperature of the removable fermentation management area is 2-7°C higher than the upper limit of the target core temperature, the temperature difference between the stack surface and the core reaches 5-12°C, or the heating rate reaches 1.5-3.0°C per hour, the industrial control cabinet 16 generates a flexible removable mixing suggestion. After manual confirmation, door closure, O2 not lower than the safety threshold, torque not exceeding the torque limit protection value, and PLC safety interlock meeting the conditions, the PLC safety control module 1607 controls the flexible removable mixing mechanism corresponding to the removable fermentation management area to run at 5-20 rpm for 30-90 seconds and pause for 30-120 seconds, while simultaneously opening the zoned air supply and return air dehumidification of the corresponding fermentation management area. During the flexible random mixing process, the PLC safety control module 1607 continuously reads the drive shaft torque, motor current, limit switch, door status, core temperature, and O2 / CO2 data. When the torque exceeds the torque limit protection value, the door is not closed, the O2 is lower than the safety threshold, or the core temperature does not decrease within 2-4 consecutive sampling cycles, the mixing is stopped and the alarm reset unit is triggered. After each batch, the target process-quality parameter table, index weight table, or control recommendation mapping table are calibrated based on the moisture content at the warehouse, sensory evaluation, physicochemical quality inspection results, sensor curves, recommended actions, manual confirmation records, and PLC execution results. The calibrated items are used for subsequent batches after manual review and PLC safety threshold verification, and do not replace the PLC safety threshold.

[0058] During the later stages of fermentation, a release verification signal is generated by combining FQI, MAI, RSI, color images, near-infrared maturity characteristics, tea leaf moisture content, and manual sensory records. When FQI reaches 80-100, MAI is between 50-85, and RSI is between 0-40, the PLC safety control module outputs a release permission prompt. This release permission prompt serves as a manual release verification signal and does not directly deactivate the PLC safety interlock or directly drive the actuator. When FQI reaches 80-100 but RSI is between 40-100, the PLC safety control module does not output a release permission prompt and maintains at least one prohibition signal from the water supply valve, heating and dehumidification module, or stirring driver / inverter.

[0059] Example 3 The present invention also provides a specific embodiment of using the multimodal microecological directional fermentation chamber for Liubao tea in Example 1 for trial batch fermentation: 1. Load 200kg of wet Liubao tea leaves into 16 compartments of a four-layer material cart. Each compartment measures 2200mm*1200mm*1800mm. Each layer has four composite sensor points, each containing a surface depth sensor 121 and a core depth sensor 122. The surface depth is 20-50mm from the surface of the tea leaves, and the core depth is 150-300mm from the surface of the tea leaves or 1 / 2-2 / 3 of the material layer thickness.

[0060] The system collects temperature, humidity, moisture, O2, CO2, VOC and pressure difference every 5 minutes, and images or near-infrared data every 30 minutes.

[0061] 2. Hardware linkage structure of industrial control cabinet The industrial control cabinet 16 has the following configurations: Input side: sensor input terminal block, image / near-infrared acquisition interface 1602, and microecological / quality inspection data interface 1603; Calculation side: industrial calculation unit 1604, process parameter and table entry storage module 1605, and FQI / MAI / RSI calculation and display module 1606; Safety interlock side: PLC safety control module 1607, safety relay and emergency stop circuit 1608, and manual confirmation and alarm reset unit 1613; Execution side: zoned air valve drive output module 1609, nozzle / metering pump drive output module 1610, agitator driver / frequency converter 1611, and execution status feedback terminal block 1612. The industrial calculation unit 1604 only generates zoned control suggestions, and the PLC safety control module 1607 executes the corresponding action when both the manual confirmation signal and the safety interlock conditions are met.

[0062] A waterproof junction box is installed on the outer wall of the insulated fermentation chamber 1, and an inlet or terminal block is installed on the side of the industrial control cabinet 16. The two are connected by a cable tray, conduit, or cable bundle to form a waterproof junction box and cable tray assembly between the chamber and the control cabinet. The signal lines, drive lines, feedback lines, and safety interlock lines of the composite sensor array, the zoned electric air valve 7, the zoned micro-mist nozzle 11, the heating and dehumidification module 15, and the flexible random stirring mechanism 23 are collected by this assembly and connected to the sensor input terminal block 1601, the zoned air valve drive output module 1609, the nozzle / metering pump drive output module 1610, the stirring driver / frequency converter 1611, and the execution status feedback terminal block 1612.

[0063] 3. Target Process - Quality Parameter Table The target process-quality parameter table records parameters by name, target lower limit (L), target upper limit (U), allowable lower limit (AL), allowable upper limit (AU), safety lower limit (SL), and safety upper limit (SU). For example, for core temperature, L is 45℃, U is 58℃, AL is 42℃, AU is 62℃, SL is 38℃, and SU is 65℃; for tea leaf moisture content, L is 28%, U is 38%, AL is 25%, AU is 42%, SL is 22%, and SU is 45%. Before use for different batches, the table entries are manually reviewed, and the PLC safety threshold verification ensures that the entries do not exceed safety boundaries.

[0064] 4. Examples of FQI, MAI, and RSI calculations FQI = 100 * (0.25C + 0.20A + 0.20T + 0.15M + 0.10N + 0.10Q), with each component normalized to 0-1 and the sum of each weight being 1. C, T, M, and N are all treated as target interval suitability parameters: C is obtained by comparing the L*a*b* value, RGB value, or color difference value after whiteboard correction with the target color interval; T is obtained by comparing the near-infrared characteristic band or principal component score with the target taste calibration interval; M is obtained by comparing the moisture content of the tea base with the target moisture interval; and N is obtained by comparing the near-infrared maturity characteristic value with the target maturity interval. When the above parameter x is within the target interval [L,U], the score is 1; when it is within [AL,L), the score is calculated as (x-AL) / (L-AL); when it is within (U,AU], the score is calculated as (AU-x) / (AU-U); and when it is within [SL,AL) or (AU,SU], the score is 0 and a manual confirmation prompt is triggered.

[0065] A is calculated using the weighted distance dA between the current VOC vector X and the target VOC vector Tg, where dA = s qrt (sum(k a *(X a -T ga )^2)), A=1-min(dA / dA max ,1), where dA max >0, VOC vector X and target VOC vector T g Including the same number and in the same order of VOC components, X a and T ga Let sum(k) be the current value and the target value of the a-th VOC component, respectively. a *(X a -T ga )^2) is the summation over each VOC component a; when dA max When =0, A will not automatically execute the judgment and will trigger manual review. Q compares the moisture, color, VOC, and near-infrared characteristic vector Y of the current batch with the average value vector Y of similar qualified batches. g The weighted distance dQ is converted, dQ=sqrt (sum(k q *(Y q -Y gq )^2)), Q=1-min(dQ / dQ max ,1); Where dQ max >0, Y and Y g Including the same number and consistent order of historical quality inspection features, Y q and Y gq Let sum(k) be the current value and mean of the q-th historical quality inspection feature, respectively. q *(Y q -Y gq )^2) is the summation of each historical quality inspection characteristic q; when dQ max When the sum of the values ​​is 0 or there are no historical samples of similar qualified batches, Q is replaced by Q=(0.25C+0.20A+0.20T+0.15M+0.10N) / (0.25+0.20+0.20+0.15+0.10), and the denominator is greater than 0; if the sum of the corresponding weights is 0, Q is calculated with equal weights for C, A, T, M and N. MAI = 100 * (0.20R + 0.18W + 0.18O + 0.18G + 0.16V + 0.10B), where each component is normalized to 0-1 and the sum of each weight is 1; R, W, O, G, and V are converted to 0-1 according to the target range, allowable range, or unit time change threshold in the target process-quality parameter table, and B is converted to 0-1 according to the accessed verification data relative to the corresponding verification threshold or qualified range; when B is not accessed, the real-time calculation weight of item B is reset to 0. Under the condition that the sum of the original weights of R, W, O, G, and V is greater than 0, the original weight of item B 0.10 is normalized and allocated according to the original weight ratio of R, W, O, G, and V, and calculated as MAI = 100 * (v1'R + v2'W + v3'O + v4'G + v5'V); when the sum of the original weights of R, W, O, G, and V is 0, MAI is calculated according to the five items with equal weights. RSI=100*max(E,H,L r G c V r (,S,D), where E, H, L r G c and V r The maximum value of each individual risk is taken. Each individual risk is divided into high-value risk type, low-value risk type and range deviation type according to the risk direction of the parameter and converted to a risk value of 0-1. The high-value risk type is used for CO2 accumulation risk and overheating upper limit risk, the low-value risk type is used for hypoxia risk, and the range deviation type is used for moisture, VOC deviation or temperature difference deviation risk. S and D are normalized to 0-1 according to the abnormal duration period or alarm status.

[0066] 5. Example of a Control Recommendation Mapping Table Project The control suggestion mapping table uses input conditions, risk level, execution output fields, PLC allowable conditions, and prohibitable execution conditions fields. The execution output fields include target actuator and execution parameter fields. Example 1: Input condition is moisture content below the target lower limit and a moisture difference between adjacent zones of 2-4 percentage points; target actuator is the corresponding zone's water supply valve or metering pump; execution parameters are water supply volume and number of water supply cycles. Example 2: Input condition is a CO2 concentration difference of 1000-3000 ppm or an O2 concentration difference of 1-3 percentage points; target actuator is the corresponding zone's air valve and dehumidification channel; execution parameters are air valve opening and dehumidification duration. Example 3: Input condition is that the core temperature, surface / core temperature difference, or heating rate reaches superheating conditions; target actuator is the flexible bulk agitator in the removable fermentation management area, linked to air supply and dehumidification. The above zone control suggestions first enter the zone control suggestion display area. Only when the signal is manually confirmed as valid and the PLC safety interlock is met will the PLC safety control module convert it into an execution output.

[0067] 6. Processing of the execution mechanism field configured by partition For the removable-stack fermentation management area, the target actuator field of the control recommendation mapping table includes the agitator / frequency converter; for the non-removable-stack fermentation management area, the target actuator field includes zoned air valves, return air dehumidification channels, water supply valves, or metering pumps, but does not include the agitator / frequency converter. When the core temperature rises or the rate of temperature rise increases in the non-removable-stack fermentation management area, the system outputs recommendations for zoned air supply, dehumidification, stopping water supply, or manual inspection. The system also limits the intensity of automatic execution when the RSI reaches 60-80, and prohibits the corresponding automatic execution action when the RSI is greater than 80, thereby preventing the erroneous generation or execution of agitation outputs in areas where flexible removable-stack agitators are not installed.

[0068] 7. Localized high CO2 levels and hypoxia treatment When CO2 levels in zones A2 and A4 continuously rise while O2 levels decrease, and the CO2 concentration difference with adjacent zones reaches 1000-3000 ppm or the O2 concentration difference reaches 1-3 percentage points, the system determines that local ventilation is insufficient. The industrial control cabinet generates a suggestion to open the corresponding zone's electric air valve and extend the dehumidification time by 8-15 minutes; the PLC is prohibited from automatically executing this suggestion if the air valve malfunctions, communication fails, or the door is not closed.

[0069] 8. Flexible scattered reactor core treatment for overheating When the core temperature of a certain removable composting management area is 2-7°C higher than the target core temperature limit, the temperature difference between the compost surface and the core reaches 5-12°C, or the heating rate reaches 1.5-3.0°C per hour, the industrial control cabinet generates a flexible removable composting agitation suggestion. After the operator confirms that the door is closed, O2 is not lower than the safety threshold, torque does not exceed the torque limit protection value, and the PLC safety interlock meets the conditions, the PLC drives the flexible removable composting agitation mechanism of the removable composting management area to run at 5-20 rpm for 30-90 seconds, pause for 30-120 seconds, the torque limit protection device limits the torque to no more than 5 N·m, and the position limit device limits the insertion depth to no more than two-thirds of the material layer thickness; at the same time, short-term zoned air supply and dehumidification are activated.

[0070] 9. Sensor fault tolerance When a sensor reading in a certain zone suddenly changes without corresponding changes in adjacent zones, other depths in the same layer, or image features, the system marks the data as potentially abnormal, reduces its weight, or triggers manual verification, and does not directly perform water replenishment or dehumidification actions based on this data. If the same sensor shows abnormalities for multiple consecutive sampling cycles, the RSI adds a sensor failure risk item, and the PLC restricts related automatic actions.

[0071] 10. Table Entry Calibration and Version Tracking After fermentation, the quality inspection results, sensory scores, moisture content at the outlet, liquor color, aroma evaluation, sensor curves, recommended actions, manual confirmation records, and PLC execution results are imported into the data interface. The system generates calibration samples on a batch-by-batch basis for calibrating the target process-quality parameter table, indicator weight table, or control recommendation mapping table. The calibrated entries undergo manual review and PLC safety threshold verification before the next batch is used. This calibration process does not replace emergency stop, over-temperature, over-humidity, oxygen deficiency, door unclosed, and drive shaft overload interlocks.

[0072] 11. Methods for verifying technical effects During verification, under the same tea batch, initial moisture content, loading amount, and sampling cycle, three groups were set up: a manual turning group, a hard stirring group, and a flexible loose-pile group based on the present invention. The time required for the core temperature to drop from its peak to a safe range, the cooling rate per unit time, the broken tea rate measured by sieving, the differences in moisture and color between different zones, and the FQI compliance rate were recorded. The broken tea rate was determined as the proportion of sieved material before and after treatment to the total sample mass; temperature and humidity deviations were calculated as the difference between the maximum and minimum values ​​of the sensors in each zone at the same time; the FQI compliance rate was calculated as the proportion of batches leaving the warehouse whose FQI reached the target range; the control suggestion consistency rate was calculated as the ratio of system-generated suggestions to manually reviewed and approved suggestions; the manual review pass rate was calculated as the ratio of the number of manually confirmed passes to the total number of manual reviews; the feedback signal integrity rate was calculated as the ratio of the number of actions that obtained feedback on opening / closing, opening degree, flow rate, speed, torque, or alarm status after the execution of the action to the total number of actions executed; and the action suppression records were calculated as the number of suggestions that were not converted into execution output due to sensor abnormalities, PLC safety thresholds, or prohibited execution conditions.

[0073] The above description of the disclosed embodiments enables those skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multimodal microecological directional fermentation chamber for Liubao tea, characterized in that, include: The insulated fermentation chamber (1) and the multi-layer partitioned material frame (4), partitioned environmental adjustment mechanism, composite sensing mechanism, flexible random mixing mechanism (23) and industrial control cabinet (16) are set in the insulated fermentation chamber. The multi-layer partitioned material frame (4) is used to divide the fermentation space of the tea blank into multiple fermentation management zones. Each fermentation management zone is equipped with the partitioned environment adjustment mechanism and the composite sensing mechanism. The fermentation management zone equipped with the flexible loose-pile stirring mechanism is designated as the loose-pile fermentation management zone (242), and the fermentation management zone without the flexible loose-pile stirring mechanism is designated as the non-loose-pile fermentation management zone (243). The partitioned environmental control mechanism is used to independently regulate the air supply, dehumidification, water replenishment, or heating of each fermentation management zone. The composite sensing mechanism is used to collect fermentation status data of Liubao tea in the corresponding fermentation management area; The fermentation status data includes: temperature, humidity, gas composition, images, and near-infrared spectroscopy. The industrial control cabinet (16) is electrically connected to the partitioned environmental control mechanism, the composite sensing mechanism and the flexible random stirring mechanism (23) respectively; The industrial control cabinet (16) is configured to: calculate the FQI quality proximity index, MAI microecological activity index and RSI abnormal risk index based on the sensing data, and generate zonal control suggestions by querying the control suggestion mapping table; After the partition control suggestion is manually confirmed and meets the PLC safety interlock conditions, it is converted to execute the corresponding partition environment adjustment mechanism or flexible random mixing mechanism (23) action, and the execution result and status feedback are recorded.

2. The multimodal microecological directional fermentation chamber for Liubao tea according to claim 1, characterized in that, The zoned environmental control mechanism includes: a zoned air supply and dehumidification system and a zoned micro-mist water replenishment system; The zoned air supply and dehumidification system includes: a main air supply pipe (5), zoned air supply branches (6) corresponding to each fermentation management zone, zoned electric air valves (7), return air dehumidification channels (8), and dehumidification outlets (9); The main air supply pipe (5) is branched to multiple zone air supply branches (6), and each zone air supply branch (6) corresponds to a fermentation management zone. The partition electric air valve (7) is installed on each partition air supply branch (6); The return air dehumidification channel (8) is located on the return air side and connected to the dehumidification port (9); The zoned micro-mist water supply system includes: a micro-mist water supply main pipe (10) and zoned micro-mist nozzles (11) corresponding to each fermentation management zone. The main water supply pipe (10) is connected to the corresponding micro-mist nozzle (11) of each fermentation management area. The main water supply pipe (10) for micro-mist is equipped with a water supply valve or a metering pump to control the on / off flow or flow rate of water supply to the micro-mist nozzles (11) in the partition.

3. The multimodal microecological directional fermentation chamber for Liubao tea according to claim 1, characterized in that, The composite sensing mechanism includes: a composite sensor array (12), an image acquisition unit, a near-infrared acquisition unit, and a weighing base; The composite sensor array includes: a surface depth sensor (121) and a core depth sensor (122). The surface depth sensor (121) and the core depth sensor (122) are used to collect data on temperature, relative humidity, tea leaf moisture, oxygen, carbon dioxide, volatile organic compounds and pressure difference of the surface and core of the fermented tea pile, respectively. The image acquisition device is used to collect the surface color, material layer flatness, and clumping state of the tea leaves; The near-infrared collector is used to collect the near-infrared reflection characteristics and near-infrared principal component scores of the tea blank. The weighing base (14) is used to collect the weight data of the tea blanks; The surface depth sensor (121) is located 20-50 mm from the surface of the tea blank, and the core depth sensor (122) is located 150-300 mm from the surface of the tea blank or at 1 / 2-2 / 3 of the thickness of the material layer; each layer of the material frame is equipped with no less than 4 sets of sensor points.

4. The multimodal microecological directional fermentation chamber for Liubao tea according to claim 1, characterized in that, The flexible bulk mixing mechanism includes: a deceleration drive assembly (239), a low-speed drive shaft (231), a flexible rounded corner paddle (232), a torque limiting protection component (233), and a position limiting component (234). The speed reduction drive assembly (239) includes a drive motor and a speed reducer, which are located outside the partitioned material frame of the dispersible fermentation management area (242) and fixed on the side wall (4) of the partitioned material frame or on the mounting bracket connected to the partitioned material frame (4). The output end of the drive motor is connected to the input end of the reducer, and the output end of the reducer is connected to one end of the low-speed drive shaft (231) through the torque limiting protection component (233). The low-speed drive shaft (231) is arranged laterally along the partitioned material frame (4) and extends into or above the tea blank layer; The multiple flexible rounded corner paddles (232) are fixedly connected to the shaft segment of the low-speed drive shaft (231) located in the partitioned material frame, and are arranged axially and circumferentially staggered along the low-speed drive shaft (231) so that when the low-speed drive shaft (231) rotates, it drives the flexible rounded corner paddles (232) to flip the tea blanks. The position limiting member (234) is provided on the side wall of the partitioned material frame (4) or on the mounting bracket of the low-speed drive shaft (231), and is limited and cooperated with the mounting part of the low-speed drive shaft (231); The rotational speed range of the low-speed drive shaft (231) is 2-30 rpm, the protection torque range of the torque limiting protection component (233) is 0.5-8 N·m, and the position limiting component (234) is used to limit the insertion depth of the flexible rounded corner paddle to no more than two-thirds of the material layer thickness.

5. The multimodal microecological directional fermentation chamber for Liubao tea according to claim 1, characterized in that, The industrial control cabinet includes: an industrial computing unit (1604), a PLC safety control module (1607), a process parameter and table entry storage module (1605), and a communication gateway and data traceability module (1614). The industrial computing unit (1604) is used to calculate the FQI quality proximity index, MAI microecological activity index and RSI abnormal risk index based on the sensing data, and to query the control suggestion mapping table to generate zonal control suggestions. The process parameter and table storage module (1605) is used to store the target process-quality parameter table, the index weight table, the control suggestion mapping table and the PLC safety threshold. The PLC safety control module (1607) is used to receive manual confirmation signals and perform safety interlock judgments. When the manual confirmation signal is valid and the safety interlock conditions are met, the zoning control suggestion is converted into an execution output. The PLC safety interlock has a higher priority than the zoning control suggestion output by the industrial computing unit. When the process parameter to be judged is within the safe range but exceeds the allowable range, the zoning control suggestion is not converted into an execution output, but only an alarm prompt is generated. When the process parameter to be judged exceeds the safe range, the corresponding automatic execution action is prohibited. The communication gateway and data traceability module are used to record batch data, recommended actions, manual confirmation records, PLC execution results, and execution status feedback.

6. The multimodal microecological directional fermentation chamber for Liubao tea according to claim 1, characterized in that, A waterproof junction box and cable tray assembly (24) is provided between the outer wall of the insulated fermentation chamber (1) and the industrial control cabinet (16). The waterproof junction box and cable tray assembly (24) of the chamber-control cabinet includes: an inlet terminal set on the outer wall of the heat-insulating fermentation chamber (1) and the side of the control cabinet; The signal lines, drive lines, feedback lines, and safety interlock lines of the composite sensing mechanism, zoned environmental adjustment mechanism, and flexible random mixing mechanism (23) are all connected to the industrial control cabinet (16) via the waterproof junction box and cable tray assembly (24) of the silo-control cabinet.

7. A method for controlling the multimodal microecological directional fermentation of Liubao tea, applied to the multimodal microecological directional fermentation chamber for Liubao tea as described in any one of claims 1-7, characterized in that, The method includes: S1. Real-time collection of fermentation status data of tea leaves in the corresponding fermentation management area through a composite sensing mechanism; the fermentation status data includes: temperature, humidity, gas composition, image and near-infrared data; S2. Calculate the FQI quality proximity index, MAI microecological activity index, and RSI abnormal risk index based on the perceived data. S3. Query the control suggestion mapping table and generate partition control suggestions based on the calculation results; S4. Manually confirm the proposed partition control and determine whether the PLC safety interlock conditions are met. S5. If manual confirmation is passed and the PLC safety interlock conditions are met, the zoning control suggestion is converted into an execution output, driving the corresponding zoning environmental adjustment mechanism to perform independent air supply, dehumidification, water replenishment or heating adjustment actions, or driving the flexible bulk mixing mechanism (23) to perform mixing actions in the bulk fermentation management area. S6. Record execution results and status feedback.

8. The method for controlling the multimodal microecological directional fermentation of Liubao tea according to claim 7, characterized in that, The formula for calculating FQI is as follows: FQI=100*(w1C+w2A+w3T+w4M+w5N+w6Q); Wherein, C is the color similarity, A is the VOC aroma vector similarity, T is the near-infrared taste calibration similarity, M is the tea base moisture suitability, N is the near-infrared maturity similarity, and Q is the historical quality inspection matching degree; C, T, M, and N are all target interval type suitability parameters, A is converted according to the weighted distance between the current VOC vector and the target VOC vector, and Q is converted according to the weighted distance between the current batch feature vector and the historical qualified batch mean vector; w1, w2, w3, w4, w5, and w6 are weight coefficients and their sum is 1.

9. The method for controlling the multimodal microecological directional fermentation of Liubao tea according to claim 7, characterized in that, The formula for calculating MAI is: MAI=100*(v1R+v2W+v3O+v4G+v5V+v6B); Wherein, R is the suitability of the core heating rate, W is the suitability of the water content change per unit time, O is the oxygen consumption characteristic per unit time, G is the carbon dioxide generation characteristic per unit time, V is the VOC change characteristic per unit time, and B is the microecological verification characteristic; each component is normalized to 0-1; when no verification data is available, the weight of item B is proportionally allocated to other items.

10. The method for controlling the multimodal microecological directional fermentation of Liubao tea according to claim 7, characterized in that, The formula for calculating RSI is: RSI=100*max(E,H,L r ,G c ,V r ,S,D); Where E represents overheating risk, H represents overhumidity risk, and L represents overheating risk. r For the risk of hypoxia, Gc is for the risk of CO2 accumulation, and V is for the risk of oxygen deficiency. r S represents abnormal VOC risk, D represents sensor failure risk, and D represents actuator abnormal risk. Each individual risk value is converted to 0-1 based on the degree to which the parameter deviates from the target range or safe range.