Automatic air conditioning method and system for three-dimensional tobacco stack

Abstract

The invention relates to the technical field of tobacco processing, in particular to an automatic air conditioning method and system for a three-dimensional tobacco stack.The method comprises the following steps that mildew frequency band amplitude is collected to generate gradient identification, a fermentation heat frequency band is positioned to generate an early warning instruction, oxygen concentration is extracted to generate a fan instruction, and drifting is detected to generate a nitrogen valve instruction; combining the inclination angle and the wind speed to generate an air conditioning scheme. According to the invention, through extraction of amplitude changes of microbial metabolism frequency bands and construction of difference gradient features, accurate quantification of mildew activity is realized, early recognition accuracy is improved, early warning is generated in combination with frequency band deviation frequency, abnormal response capability is enhanced, differential fan strategies are generated based on oxygen concentration and accumulation topology, and circulation efficiency is optimized. Sensor drift and humidity gradient are fused, dynamic compensation of nitrogen injection is realized, a dead angle is identified in combination with an inclination angle and a wind speed, a self-adaptive air conditioning scheme is generated, and an efficient closed-loop system from early warning to regulation is constructed.

Description

Technical Field

[0001] The present invention relates to the technical field of tobacco processing, and in particular to an automated controlled atmosphere method and system for a three-dimensional tobacco stack. Background Art

[0002] The technical field of tobacco processing includes links such as storage, processing, and quality maintenance after tobacco leaves are harvested. The core content involves precise regulation of the internal environmental parameters of the tobacco stack to ensure the quality of tobacco leaves. Systematic technologies cover aspects such as temperature and humidity control, pest control, and control of chemical composition changes. Traditional methods rely on manual monitoring and manual adjustment, resulting in low efficiency and poor consistency. Modern technologies have gradually introduced automated equipment and sensors. However, due to the complex structure of the three-dimensional tobacco stack and uneven distribution of environmental parameters, targeted solutions are still needed.

[0003] Among them, the automated controlled atmosphere method refers to optimizing the tobacco storage environment by automatically adjusting the gas composition inside the three-dimensional tobacco stack. This technical matter covers oxygen concentration regulation, carbon dioxide concentration balance, and temperature and humidity maintenance. Gas concentration sensors are used to continuously monitor the internal environmental parameters of the tobacco stack. The input and discharge ratios of gases are adjusted through automatically controlled valves. Combined with circulation fans, gas is promoted to be evenly distributed. Preset programs are used to dynamically correct environmental parameter deviations to ensure that the gas composition is consistent in each area.

[0004] The prior art relies on preset programs and global unified control strategies, lacking effective means to characterize the dynamic characteristics of microbial metabolic activities inside the tobacco stack, resulting in a lag in mildew warning. For example, traditional methods cannot analyze the correlation between frequency band amplitude differences and metabolic activities and are difficult to quantify the activity gradient; automated equipment does not perform differential control in combination with the hierarchical characteristics of bulk density, resulting in low local gas circulation efficiency. For example, a fixed-speed fan cannot adapt to the oxygen consumption gradient of different layers; the calibration mechanism of oxygen sensors does not consider the drift accumulation effect and the spatial coupling relationship of humidity gradients, resulting in compensation delay. For example, when nitrogen injection is not preferentially triggered in high-humidity areas, local oxygen concentration imbalance exacerbates the mildew risk; the identification of dead flow areas depends on static empirical parameters and does not build a model in combination with real-time data of inclination offset and wind speed, resulting in uneven gas distribution. For example, traditional air supply strategies cannot dynamically eliminate the stagnant areas caused by the deformation of the tobacco stack. The linear control mode of the prior art is difficult to adapt to the dynamic change requirements in a complex hierarchical environment. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings existing in the prior art and propose an automated controlled atmosphere method and system for a three-dimensional tobacco stack.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: An automated controlled atmosphere method for a three-dimensional tobacco stack includes the following steps: S1: In the sealed cavity of the tobacco stack, collect the amplitude changes in the frequency bands associated with the metabolism of mildew microorganisms in the gas spectrum, compare the amplitude differences in the same frequency band of adjacent sampling periods, arrange them in the order of frequency band numbers to form a gradient feature set, and generate a mildew activity gradient identifier for the tobacco stack; S2: Based on the mildew activity gradient identifier of the tobacco stack, locate the gradient fluctuation range of the tobacco fermentation heat release frequency band, calculate the continuous deviation times of the frequency band gradient value from the safe frequency domain threshold, and generate a surface fermentation heat warning instruction; S3: Invoke the surface fermentation heat warning instruction, extract the stratified monitoring values of the oxygen concentration in the tobacco stack, combine with the attenuation trend of the mildew activity gradient, match the stratified topological structure of the stacking density, and generate a stratified rotation speed instruction for the circulation fan; S4: According to the operation feedback of the stratified rotation speed instruction of the circulation fan, detect the cumulative offset value of the stratified drift of the oxygen sensor, combine with the stratified gradient change of humidity, calculate the compensation priority, and generate a stratified opening and closing instruction for the nitrogen injection valve; S5: Based on the action coordinates of the stratified opening and closing instruction of the nitrogen injection valve, synchronously obtain the axial offset amount of the stacking inclination angle, calculate the product effect index with the stratified air supply wind speed, and generate an automated gas conditioning scheme for the tobacco stack.

[0007] As a further solution of the present invention, the mildew activity gradient identifier of the tobacco stack is specifically an amplitude sequence of associated frequency bands and a mildew activity difference gradient. The surface fermentation heat warning instruction includes a deviation frequency threshold, a safe threshold interval, and a warning level parameter. The stratified rotation speed instruction of the circulation fan specifically refers to the stratified oxygen concentration distribution value, the mildew activity attenuation gradient, and the density stratified topological coordinates. The stratified opening and closing instruction of the nitrogen injection valve includes the cumulative drift amount of the oxygen sensor, the humidity distribution gradient vector, and the compensation priority sequence. The automated gas conditioning scheme of the tobacco stack is specifically the stacking inclination angle offset parameter, the diversion dead angle coordinate mapping, and the stratified wind speed regulation coefficient.

[0008] As a further solution of the present invention, the specific steps of S1 are as follows: S101: Collect continuous-time gas fluctuation spectrum data in the sealed cavity of the tobacco stack, extract the characteristic frequency bands associated with the metabolism of mildew microorganisms, extract the amplitude parameters corresponding to the frequency band numbers, establish the correspondence between the frequency band numbers and the amplitudes, and obtain the frequency band amplitude matrix; S102: Invoke the amplitude sequences in the same frequency band of adjacent sampling periods in the frequency band amplitude matrix, calculate the absolute difference of the amplitudes in adjacent periods in the corresponding frequency band, arrange the difference values in the order of frequency band numbers, and obtain the frequency band difference sequence set; S103: According to the difference values in the frequency band difference sequence set, construct a gradient change array according to the frequency band numbers, identify the quantitative situation representing the mildew activity, and obtain the mildew activity gradient identifier of the tobacco stack.

[0009] As a further solution of the present invention, the specific steps of S2 are as follows: S201: Based on the mildew activity gradient identification of the tobacco stack, extract the frequency numbers in the corresponding frequency band of the surface tobacco, compare the numerical sequences corresponding to the numbers, analyze the continuous interval range of the fermentation heat release-related frequency band in the gradient identification, and obtain the surface frequency band gradient interval; S202: Call the gradient values corresponding to the frequency bands in the surface frequency band gradient interval, take the stored safety frequency domain threshold as the judgment benchmark, compare the offset direction between the frequency band value and the threshold, count the number of offsets occurring in the continuous period, and obtain the continuous offset times of the frequency band; S203: According to the continuous offset times of the frequency band, judge whether the set trigger condition is reached. If it is satisfied, generate a warning status code according to the three-level warning rule, and construct an instruction output with the current period as the unit to obtain the surface fermentation heat warning instruction.

[0010] As a further solution of the present invention, the specific steps of S3 are as follows: S301: Call the surface fermentation heat warning instruction, extract the monitoring values of the oxygen concentration sensors corresponding to the stratified positions in the tobacco stack, classify and organize all sensor data according to the collection time and stratified height, establish a mapping data table of oxygen concentration and stratification levels, and obtain the oxygen concentration stratified distribution value; S302: According to the oxygen concentration stratified distribution value, extract the mildew activity gradient data sequence in the corresponding time period of each layer, calculate the sequence difference, and judge the change trend of the decline rate in the continuous time period. Quantify the gradient difference trend of each layer into a percentage form to obtain the gradient decline rate trend value; S303: Combine the gradient decline rate trend value and the oxygen concentration stratified distribution value, match the spatial coordinate system in the tobacco stacking density topology structure, calculate the ventilation control error characteristic value at each coordinate level, and obtain the hierarchical rotation speed instruction of the circulation fan.

[0011] As a further solution of the present invention, the specific calculation formula of the ventilation control error characteristic value at each coordinate level is as follows: ; Wherein, represents the ventilation control error characteristic value in the spatial coordinates of the i-th row, j-th column and k-th layer, represents the weighted average value of the gradient decline rate trend values at the coordinate points of the i-th row and j-th column, represents the standard deviation of the oxygen concentration stratified distribution value at the coordinate points of the i-th row and j-th column, represents the tobacco stacking density value corresponding to the spatial position of the i-th row, j-th column and k-th layer, represents the gradient of the flow path change in the coordinate area of the i-th row and j-th column, Represents the initial fan control template value corresponding to the coordinate of the i-th row, j-th column, and k-th layer.

[0012] As a further solution of the present invention, the specific steps of S4 are: S401: extracting the monitoring value of the stratified oxygen sensor according to the operation feedback signal of the stratified speed command of the circulating fan, calculating the difference in the continuous cycle according to the time series, and performing cumulative difference statistics with the initial reference value to obtain the cumulative amplitude of oxygen drift; S402: extracting the layered humidity monitoring value in the stack according to the accumulated amplitude of oxygen drift, calculating the humidity difference between the layers, and determining the gradient change direction, corresponding to the drift amplitude according to the layer number, and obtaining the humidity offset matching direction value; S403: calling the humidity offset matching direction value, calculating the joint change intensity index of the oxygen drift amplitude and the humidity direction, generating the corresponding layered opening and closing state code according to the set priority judgment threshold, and obtaining the nitrogen injection valve layered opening and closing instructions.

[0013] As a further solution of the present invention, the calculation formula of the combined change intensity index of the oxygen drift amplitude and the humidity direction is specifically: ; in, An index representing the combined change intensity of oxygen drift amplitude and humidity direction, represents the drift amplitude of oxygen concentration measured for the jth time, represents the humidity value measured for the jth time, represents the humidity direction value measured for the jth time, represents the reference humidity direction value of the jth measurement, Represents the total number of measurements.

[0014] As a further solution of the present invention, the specific steps of S5 are: S501: based on the coordinates of the action area of ​​the nitrogen injection valve layered opening and closing instructions, synchronously obtain the real-time monitoring value of the inclination sensor in the corresponding space, extract the offset data in the axial dimension, and collect them according to the layer number to obtain the axial inclination offset; S502: According to the axial inclination angle offset, the wind speed monitoring data of the air supply duct in the area is matched, the inclination angle value and the wind speed value consistent with the coordinates of the diversion dead angle area are extracted, and the two are multiplied according to the coordinate pair to obtain the airflow influence of the diversion dead angle; S503: Call the airflow influence amount of the diversion dead corner, combine the coordinate distribution range of the diversion point, assemble the wind speed product value by area and generate the ventilation adjustment instruction structure of the corresponding position, establish a numerical configuration set for regulating the circulation channel, and obtain the automatic air conditioning plan of the chimney.

[0015] An automated controlled atmosphere system for a three-dimensional cigarette stack, comprising: The spectrum recognition module collects the gas fluctuation spectrum amplitude in the frequency band associated with the metabolism of mildew microorganisms in the sealed cavity of the cigarette stack, compares the amplitude differences in the same frequency band within adjacent sampling periods, arranges the difference values according to the frequency band serial numbers to form a continuous gradient set, screens the number of frequency bands with amplitude changes, and obtains the mildew activity gradient identifier of the cigarette stack; The heat warning module, based on the mildew activity gradient identifier of the cigarette stack, extracts the gradient fluctuation value in the surface heat release frequency band, counts the number of times of continuously deviating from the stored safety frequency domain threshold, and determines whether the number of deviations reaches the preset trigger condition to obtain the surface fermentation heat warning instruction; The wind speed instruction module calls the surface fermentation heat warning instruction, extracts the real-time monitoring value of the oxygen concentration with different layers of the cigarette stack, combines the downward trend of the mildew activity gradient identifier, and matches the spatial coordinates of the corresponding stacking density layers to obtain the hierarchical rotation speed instruction of the circulation fan; The nitrogen injection control module, according to the feedback signal of the hierarchical rotation speed instruction of the circulation fan, detects the cumulative offset value of the drift of the corresponding layer oxygen sensor, extracts the humidity distribution gradient direction, calculates the cross-change intensity of the two, and obtains the hierarchical opening and closing instruction of the nitrogen injection valve; The scheme generation module, based on the action area of the hierarchical opening and closing instruction of the nitrogen injection valve, obtains the axial offset of the regional stacking inclination monitoring value, extracts the wind speed value of the air supply pipeline, calculates the product effect of the inclination offset and the wind speed, and obtains the automated controlled atmosphere scheme of the cigarette stack.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, by extracting the amplitude change in the microbial metabolism frequency band, constructing the differential gradient feature, realizing the accurate quantification of mildew activity, improving the early recognition accuracy, generating a warning by combining the frequency band deviation frequency, enhancing the abnormal response ability, generating a differential fan strategy based on the oxygen concentration and the stacking topology, optimizing the circulation efficiency, integrating the sensor drift and the humidity gradient, realizing the dynamic compensation of nitrogen injection, combining the inclination and the wind speed to identify the dead angle, generating an adaptive controlled atmosphere scheme, and constructing an efficient closed-loop system from warning to regulation. Description of the Drawings

[0017] Figure 1 It is a schematic diagram of the step flow of the present invention; Figure 2 It is a system module diagram of the present invention. Detailed Embodiment

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation to the present invention. In addition, in the description of the present invention, the meaning of "a plurality of" is two or more, unless otherwise specifically defined.

[0020] Please refer to Figure 1 , an automated controlled atmosphere method for a three-dimensional tobacco stack, comprising the following steps: S1: In the sealed cavity of the tobacco stack, collect the amplitude change of the frequency band strongly associated with the metabolism of tobacco mildew microorganisms in the gas fluctuation spectrum, compare the absolute value of the difference in the amplitude of the same frequency band in adjacent sampling periods, arrange them in the order of frequency band numbers as a gradient feature set, and generate a mildew activity gradient identifier for the tobacco stack; S2: Based on the mildew activity gradient identifier of the tobacco stack, locate the gradient fluctuation range of the fermentation heat release frequency band on the surface layer of the tobacco, calculate the continuous deviation times between the frequency band gradient value and the stored safety frequency domain threshold. When the deviation times reach the preset trigger condition, generate a surface layer fermentation heat warning instruction; S3: Invoke the surface layer fermentation heat warning instruction, extract the real-time monitoring values of the oxygen concentration distribution in the internal layers of the tobacco stack, combine the attenuation trend of the mildew activity gradient decline rate, match the spatial coordinates of the tobacco stacking density stratified topology, and generate a hierarchical rotation speed instruction for the circulation fan; S4: According to the operation feedback signal of the hierarchical rotation speed instruction of the circulation fan, detect the cumulative offset amplitude of the stratified drift amount of the oxygen sensor of the tobacco stack, combine the gradient change direction of the humidity stratified distribution of the tobacco stack, calculate the trigger priority of the compensation requirement, and generate a hierarchical opening and closing instruction for the nitrogen injection valve; S5: Based on the coordinate of the action area of the hierarchical opening and closing instruction of the nitrogen injection valve, synchronously obtain the axial offset amount of the real-time monitoring value of the stacking inclination angle of the tobacco stack, calculate the product effect of the inclination angle offset amount of the diversion dead angle coordinate and the wind speed of the stratified air supply pipeline, and generate an automated controlled atmosphere scheme for the tobacco stack.

[0021] The mildew activity gradient identification of the tobacco stack specifically includes the amplitude sequence of the associated frequency band and the mildew activity difference gradient. The surface fermentation heat warning instruction includes the deviation frequency threshold, the safety threshold interval, and the warning level parameter. The hierarchical rotation speed instruction of the circulation fan specifically refers to the hierarchical oxygen concentration distribution value, the mildew activity attenuation gradient, and the density hierarchical topological coordinate. The hierarchical opening and closing instruction of the nitrogen injection valve includes the drift accumulation amount of the oxygen sensor, the humidity distribution gradient vector, and the compensation priority sequence. The automated gas conditioning scheme for the tobacco stack specifically includes the stacking inclination offset parameter, the diversion dead angle coordinate mapping, and the hierarchical wind speed regulation coefficient.

[0022] The specific steps of S1 are as follows:

[0023] S101: Collect the gas fluctuation spectrum data in the sealed cavity of the tobacco stack for a continuous period, extract the characteristic frequency bands associated with the metabolism of mildew microorganisms, extract the amplitude parameters corresponding to the frequency band numbers, establish the correspondence between the frequency band numbers and the amplitudes, and obtain the frequency band amplitude matrix; First, a spectrum acquisition device needs to be installed. This device should have a high-sensitivity detection ability for weak gas fluctuations. The acquisition time period is set to collect once every 10 minutes, and the acquisition duration each time is 60 seconds. The sampling duration is not less than 24 hours to cover the periodic change process of microbial metabolism. Sampling 10 times per second can obtain 600 groups of fluctuation data. Subsequently, perform frequency domain conversion on these 600 groups of data, extract the frequency band data in the range from 0.5 Hz to 10 Hz, and divide it into several sub-frequency bands. The division method is that the width of each frequency band is set to 0.5 Hz, with a total of 19 frequency bands, numbered from f1 to f19 respectively. Statistically analyze the signal amplitudes of the frequency points collected within each frequency band, and calculate the average amplitude of each point within the frequency band. For example, the amplitudes of the 10 frequency points sampled in frequency band f5 are 0.12, 0.13, 0.14, 0.13, 0.12, 0.15, 0.14, 0.13, 0.13, 0.12, then the amplitude of f5 can be recorded as 0.131. By this method, calculate the amplitudes of each frequency band at different time points respectively, forming a two-dimensional matrix structure of time × frequency band. Each row corresponds to a time point, and each column corresponds to the average amplitude of a frequency band. The finally obtained frequency band amplitude matrix can be used for subsequent fluctuation trend analysis. The data structure of this matrix can be improved by gradually filling the matrix elements. In actual operation, 6 groups of sampling data will be formed per hour, and 144 groups of data will be formed in 24 hours. The final frequency band amplitude matrix is a matrix structure of 144 rows × 19 columns, ensuring complete coverage of the metabolic activities of mildew microorganisms.

[0024] S102: Call the amplitude sequences of the same frequency band in adjacent sampling periods in the frequency band amplitude matrix, calculate the absolute difference of the amplitudes in adjacent periods corresponding to the frequency band, arrange the difference values according to the frequency band numbers, and obtain the frequency band difference sequence set; For the data structure in the frequency band amplitude matrix, the amplitude data corresponding to the same frequency band number at two adjacent sampling times are respectively extracted for difference calculation. For example, at the 1st moment and the 2nd moment, the amplitudes of the frequency band f7 are 0.145 and 0.151 respectively, and their difference is 0.006. The calculation of this difference is performed once for all frequency bands to complete the extraction of the complete frequency band difference data set between one sampling period. The absolute value of this difference should be taken to exclude the interference of the rising or falling direction on the analysis. The differences of all frequency bands form a difference sequence, such as [f1 difference, f2 difference,..., f19 difference]. To facilitate the identification of mildew-related fluctuation characteristics, an identification threshold needs to be set for these differences. The setting method is as follows: In the initial sampling stage, the value range of all differences of all frequency bands within the first 12 hours is statistically analyzed and arranged in ascending order. The difference value at the 90% position in the sorting result is taken as the threshold. For example, if the value at the 90% position among all differences is 0.025, then 0.025 is set as the current difference identification threshold. Any frequency band difference greater than or equal to 0.025 is marked as an active change frequency band. If within a certain time period, the differences of f4, f8, and f11 are 0.028, 0.031, and 0.035 respectively, all of which are higher than this threshold, then they are included in the difference frequency band sequence. This sequence will be used to analyze the change trends and their mutation behaviors between different frequency bands in the subsequent analysis, and further determine the mildew activity fluctuation range.

[0025] S103: According to the difference values in the frequency band difference sequence set, construct a gradient change array according to the frequency band number, identify the quantitative situation representing mildew activity, and obtain the mildew activity gradient identifier of the tobacco stack; After obtaining the differential frequency band sequences for each time period, in order to further analyze the fluctuation trend of the frequency band differences, it is necessary to calculate the gradient change, that is, to analyze the increase and decrease of the difference values between each frequency band. Perform a difference operation on two adjacent frequency bands in a difference sequence. For example, the difference of f3 is 0.020, the difference of f4 is 0.028, and the difference of f5 is 0.023. Then the corresponding gradient changes are 0.028 minus 0.020 which is 0.008, and 0.023 minus 0.028 which is -0.005. This operation is sequentially performed on the frequency band difference values within each time period to form a gradient change sequence. The length of the gradient sequence is the number of frequency bands minus 1. That is, when the total number of frequency bands is 19, the length of each gradient sequence is 18. To identify the gradient fluctuation degree of the mutant frequency bands, it is necessary to set a gradient change threshold. The threshold setting rule is: among the absolute values of all gradient values, statistically calculate the value range of all gradients within the first 12 hours, calculate its average value and standard deviation, and use "the average value plus 1.5 times the standard deviation" as the threshold reference. If the gradient value of a certain frequency band is greater than this reference value, it is marked as a frequency band mutation point. For example, if the gradient value of a certain frequency band is 0.011, the historical gradient average value is 0.006, and the standard deviation is 0.003, then the reference value is 0.006 + 0.0045 = 0.0105. Since this gradient value is 0.011, which is higher than the reference value, this frequency band is marked as a mutation point. All the marked frequency bands and their gradient values are constructed into a frequency band gradient identification set, and this identification set will be used as a key reference index in the subsequent mildew identification model.

[0026] The specific steps of S2 are as follows:

[0027] S201: Based on the gradient identification of the mildew activity of the tobacco stack, extract the frequency numbers corresponding to the surface tobacco in the corresponding frequency bands, and compare the numerical sequences corresponding to the numbers to analyze the continuous interval range of the fermentation heat release-related frequency bands in the gradient identification, and obtain the surface frequency band gradient interval; First, filter the frequency band range corresponding to the surface tobacco in the dataset. The spectral response of the surface tobacco gas is usually concentrated between 2.5 Hz and 6.0 Hz. Therefore, the corresponding frequency band numbers are from f6 to f13. It is necessary to sequentially extract the frequency numbers corresponding to f6 to f13, and determine the sub-frequency point numbers according to the sampling frequency in each frequency band. For example, the frequency point numbers included in f6 are from f6_1 to f6_5. After obtaining the frequency numbers, the amplitude data corresponding to the corresponding numbers at different sampling times are extracted one by one to form a frequency point sequence, and then these numerical sequences are compared longitudinally in the order of the frequency points. The comparison operation refers to performing a difference process on the amplitudes of each frequency point at different time periods. If the amplitudes of a certain frequency point in three consecutive cycles are 0.025, 0.028, and 0.031 in sequence, then its change amplitude increases by 0.003 per cycle. If this change shows a similar trend at multiple frequency points, it is judged that this frequency band has a continuous growth trend. Subsequently, filter these frequency bands with obvious amplitude increase, and extract the frequency band number sections where the gradient values in the gradient identifier are continuously positive or continuously greater than the set gradient reference value. The set positive gradient reference value is 0.006, which is obtained from the average value of the gradient change value historical data within a 72-hour window plus a correction coefficient of 0.002. If the gradient values of three consecutive frequency bands are 0.007, 0.009, and 0.008 respectively, all greater than 0.006 and in the same direction, then these three frequency bands form a group of gradient continuous intervals. Combining the frequency band number range and numerical judgment, it is finally confirmed that f8 to f10 form a continuous interval, forming the final surface frequency band gradient interval.

[0028] S202: Call the gradient value corresponding to the frequency band in the surface frequency band gradient interval, compare the offset direction between the frequency band value and the threshold based on the stored safe frequency domain threshold as the judgment benchmark, count the number of offsets occurring in consecutive cycles, and obtain the continuous offset times of the frequency band; Extract the gradient value data of all frequency bands within this frequency band range in each cycle in sequence. For example, if the surface frequency band range is from f8 to f10, then it is necessary to extract all the gradient values of f8, f9, and f10 in the 1st to the nth cycles respectively. Subsequently, for each frequency band, the gradient value in each cycle is compared with the set storage safety frequency domain threshold, which is set to ±0.010 according to the tobacco storage safety experience standard. Determine whether the frequency band value deviates from this threshold range. When performing the comparison operation, first determine whether the gradient value is higher than +0.010 or lower than -0.010. If either condition is met, it is recorded as one offset. For example, if the gradient values of the f9 frequency band in the 5th to the 8th cycles are 0.012, 0.011, 0.014, and 0.013, and all four consecutive cycles are greater than 0.010, then the number of consecutive offsets is 4 times. On the contrary, if the gradient value in any one of the cycles is 0.009, the continuity is interrupted and the counting restarts from the next cycle. During the entire sampling cycle, it is necessary to count the total number of consecutive offsets for each frequency band. This statistical operation needs to be completed separately for different frequency bands. The final output format is a pair of the frequency band number and its corresponding number of consecutive offsets. For example, f8 has 3 offsets, f9 has 4 offsets, and f10 has 2 offsets. If the number of interrupted cycles in the interval of consecutive cycles does not exceed 1 time (that is, there is 1 non-offset in the interval, but the total sum of the consecutive offset cycles before and after is greater than or equal to 3 times), it is still included in the consecutive offsets. During the judgment process, the definition of "offset" uses the threshold of ±0.010 as the standard, and this value is set based on the actual storage temperature and humidity control standard combined with the historical fluctuations of the frequency band. The historical sampling data shows that when there is no increase in fermentation heat, the maximum fluctuation of this threshold does not exceed ±0.008. To avoid misjudgment, 0.010 is set as the boundary value.

[0029] S203: According to the number of consecutive offsets of the frequency band, determine whether the set trigger condition is met. If it is satisfied, generate a warning status code according to the three-level warning rule, and construct an instruction output in units of the current cycle to obtain the surface fermentation heat warning instruction; Judging whether each frequency band meets the set trigger conditions one by one. The set three-level early warning rules are as follows: If the continuous deviation times of a certain frequency band ≥ 3 times, it is judged as a first-level early warning; if the continuous deviation times ≥ 5 times, it is judged as a second-level early warning; if the continuous deviation times ≥ 7 times, it is judged as a third-level early warning. For example, if the continuous deviation times of frequency band f9 is 4 times, it meets the first-level early warning but does not reach the second-level standard. Further, summarize the early warning levels of different frequency bands in the current sampling period to form a list of early warning state structures for each period. If f8 is at the first level, f9 is at the first level, and f10 has no early warning in the current period, then set the early warning state of this period to the first-level status code 1. The status codes are set as follows: no early warning is 0, first level is 1, second level is 2, and third level is 3. If any frequency band in the current period reaches the third-level early warning, the period status code is 3. When there are different-level early warnings for multiple frequency bands, the maximum level is used as the period status code. Subsequently, encapsulate the early warning status code of the current period and output it as an instruction structure. The instruction contains the period number, the corresponding early warning status code, the frequency band number where the deviation occurs, and the deviation times, etc. The format can be set as [period number, status code, [(frequency band number 1, deviation times), (frequency band number 2, deviation times),...]]. For example, if the current period number is 120, the status code is 2, the deviation times of frequency band f7 is 6, and f9 is 5, then the output is [120, 2, [(f7, 6), (f9, 5)]]. This structure will be used as a parameter for the subsequent control system to call, and finally generate a surface fermentation heat early warning instruction.

[0030] The specific steps of S3 are as follows:

[0031] S301: Call the surface fermentation heat early warning instruction, extract the monitoring values of the oxygen concentration sensors corresponding to the stratified positions in the tobacco stack, classify and organize all sensor data according to the acquisition time and the stratified height, establish a mapping data table of the oxygen concentration and the stratification level, and obtain the oxygen concentration stratified distribution value; First, determine the warning cycle time point as the key identification period, select one hour before and after this cycle time point as the data sampling interval, extract all the monitored value data of the oxygen concentration sensors installed at different height stratification positions inside the tobacco stack, perform positioning and matching according to the sensor number and the height where it is located, label the height value as the layer number. For example, assume the total height of the tobacco stack is 3.6 meters, with each layer being 0.3 meters, then it is divided into 12 layers, corresponding to the numbers L1 to L12. Obtain the oxygen concentration values at a sampling frequency of once every 10 minutes for each layer. For example, 6 values are collected in the L3 layer within this interval, which are 20.8%, 20.6%, 20.7%, 20.5%, 20.4%, and 20.3% respectively. After sorting these values in ascending order of time stamps, store them in the structure table. The content of each row of the record is [time stamp, layer number, oxygen concentration]. Then, perform hierarchical classification and grouping on all the collected data, that is, organize the data at different time points under the same layer into a group for subsequent processing. The field structure of the formed mapping data table is: layer number, corresponding sampling time period, list of oxygen concentration values. Horizontally, this data table is the layer, and vertically, it is the time dimension. The data can be stored through a three-dimensional structure array, where the third dimension is the list of oxygen concentration values. The sample structure is: L5: [10:00 → 20.4%, 10:10 → 20.3%, 10:20 → 20.2%...]. In this way, the space-time distribution data structure of the oxygen concentration of the entire tobacco stack is constructed, and finally the stratified distribution value of the oxygen concentration is obtained.

[0032] S302: According to the stratified distribution value of the oxygen concentration, extract the mildew activity gradient data sequence within the corresponding stratified time period, calculate the sequence difference, and judge the change trend of the decline rate within a continuous time period. Quantify the gradient difference trend of each layer into a percentage form to obtain the gradient decline rate trend value; Read the sampling time period ranges corresponding to each level layer by layer, and extract the mildew activity gradient data sequences corresponding to that time period. For example, the time period corresponding to layer L6 is from 10:00 to 11:00, and the gradient values in this period are 0.026, 0.023, 0.019, 0.014, 0.011, 0.008 in sequence. Then, the gradient sequence constructed under this layer is [G1, G2, G3, G4, G5, G6]. Subsequently, calculate the difference changes between adjacent time points. For example, G2−G1 is −0.003, G3−G2 is −0.004, and the difference change sequence calculated in this way is [−0.003, −0.004, −0.005, −0.003, −0.003]. Then, judge whether there is a continuous downward trend in the difference sequence. The judgment method is that three or more consecutive differences are negative and the values gradually decrease. For example, the consecutive −0.003 to −0.005 in the above sequence is a continuous downward section. Such a downward trend needs to be quantified as a percentage. The percentage is calculated as the total difference between the initial value and the end value divided by the initial value, and then multiplied by 100. For example, when 0.026 drops to 0.008, the total difference is 0.018, and the percentage is 0.018÷0.026×100≈69.23%. Finally, record the gradient descent rate of this layer as 69.23%. All levels are processed in this way. If the initial value of a certain layer is 0.030 and the final value is 0.029, the decrease amplitude is 0.001 and the decrease rate is 3.33%. Set a threshold of 5%. The decrease rate lower than this value is marked as no trend change. This threshold comes from the lower limit statistical value of the effective intervention threshold of the mildew reaction. The average minimum effective response interval of multiple storage batches is set to 0.002, and the corresponding rate lower limit is 5%. Finally, the trend values of all levels are output in the form of key-value pairs of layer numbers and rate percentages, forming a gradient descent rate trend value structure set.

[0033] S303: Combine the gradient descent rate trend value with the stratified oxygen concentration distribution value, match the spatial coordinate system in the tobacco stacking density topology, calculate the ventilation regulation error characteristic value at each coordinate level, and obtain the hierarchical rotation speed command of the circulation fan; The specific calculation formula for the ventilation regulation error characteristic value at each coordinate level is: ; Among them, represents the ventilation regulation error characteristic value in the spatial coordinates of the i-th row, j-th column, and k-th layer, represents the weighted average value of the gradient descent rate trend value at the coordinate point of the i-th row and j-th column, represents the standard deviation of the stratified oxygen concentration distribution value at the coordinate point of the i-th row and j-th column, represents the tobacco stacking density value corresponding to the spatial position of the i-th row, j-th column, and k-th layer, represents the gradient of the flow path change in the coordinate region of the i-th row and j-th column. represents the initial fan control template value corresponding to the coordinates of the i-th row, j-th column, and k-th layer.

[0034] Parameter acquisition and calculation: : By collecting the trend values of the gradient descent rate at multiple time points at the coordinate point of the -th row and -th column, and calculating their weighted average. Assume that the values collected at the time point are 0.85, 0.90, and 0.95 respectively, and the weights are 0.2, 0.3, and 0.5 respectively, then: ; ; : At the coordinate point of the -th row and -th column, collect the standard deviation of the stratified distribution value of the oxygen concentration . Assume that the oxygen concentrations in the layer are 20.8%, 20.5%, and 20.7% respectively, then: , , ; ; ; Formula calculation: Substitute the above values into the formula: ; Calculate the denominator: ; Calculate the numerator: ; Calculate the whole: ; ; The result shows that in the spatial coordinates of the -th row, -th column, and -th layer, the ventilation control error eigenvalue is 59.84, indicating that there is a large deviation between the current fan operation rate and the expected value, and adjustment is required.

[0035] The specific steps of S4 are as follows:

[0036] S401: Extract the monitored values of the stratified oxygen sensor based on the operation feedback signal of the stratified rotation speed command of the circulation fan, calculate the differences within consecutive periods in time series, and perform cumulative difference statistics with the initial reference value to obtain the cumulative amplitude of oxygen drift; First, call the execution feedback status code from the system to confirm whether the fan speed change operation is completed for each level within the specified period. After successful confirmation, extract the monitored value data of the oxygen concentration sensor at that level. Set the feedback moment as the time point t0, and collect the oxygen concentration values for each 10 - minute period starting from t0 and extending up to t0 + 1 hour at most, forming a time - series data set. For example, if the feedback of layer L4 at t0 is 10:00, then collect the oxygen concentration values for six periods: 10:10, 10:20, 10:30, 10:40, 10:50, 11:00. If the collected values are 20.6%, 20.7%, 20.8%, 20.7%, 20.9%, 21.0%, then calculate the differences between adjacent two periods. For example, 20.7% - 20.6% = +0.1%, 20.8% - 20.7% = +0.1%, and thus construct the difference sequence as [+0.1%, +0.1%, - 0.1%, +0.2%, +0.1%]. Subsequently, perform cumulative difference statistics with the initial reference value. The initial reference value is defined as the oxygen concentration value at the feedback moment t0, that is, 20.6%. Cumulatively calculate the difference between each subsequent period value and 20.6%. That is, the cumulative difference is (20.7 - 20.6)+(20.8 - 20.6)+(20.7 - 20.6)+(20.9 - 20.6)+(21.0 - 20.6)=0.1 + 0.2 + 0.1 + 0.3 + 0.4 = 1.1%. Obtain the cumulative amplitude of oxygen drift within 1 hour after feedback for this layer as +1.1%. For example, if the initial value of layer L7 is 21.0%, and the subsequent values are 21.0%, 21.1%, 20.9%, 20.8%, 20.7%, 20.5% respectively, the cumulative difference is 0.0 + 0.1 - 0.1 - 0.2 - 0.3 - 0.5=-1.0%, then the drift cumulative amplitude is - 1.0%. The same calculation process is performed for all levels, and the final output structure is a mapping set of each layer number and its drift cumulative value.

[0037] S402: Extract the monitored values of the stratified humidity in the stack according to the cumulative amplitude of oxygen drift, calculate the humidity difference between layers, and judge the direction of gradient change, corresponding to the drift amplitude according to the layer number, to obtain the humidity offset matching direction value; According to the calculation result of the cumulative amplitude of oxygen drift, extract the humidity monitoring value sequence within the corresponding cycle of the hierarchical number. Based on the same time axis, call the humidity sensor data once every 10 minutes for each layer. Similarly, starting from the feedback cycle t0, collect the data for the subsequent 6 cycles and organize them into a hierarchical humidity table structure. In the example, the humidity values of layer L4 are 75%, 76%, 77%, 76%, 75%, 74%, and those of layer L5 are 76%, 77%, 78%, 77%, 76%, 75%. Calculate the humidity difference between layer L4 and layer L5. At each cycle point, subtract L4 from L5 to obtain the result sequence [+1%, +1%, +1%, +1%, +1%, +1%]. Analyze its change direction, that is, determine whether the positive and negative signs of the difference are continuously consistent. If it is continuously positive, it is judged that the humidity rises from the lower layer to the higher layer, and the direction is marked as "rising". If it is negative, it is "falling". If the change direction reverses two or more times within the cycle, it is marked as "fluctuating". For example, the humidity values of layer L7 are 72%, 73%, 73%, 72%, 71%, 70%, and those of layer L8 are 74%, 74%, 74%, 74%, 74%, 74%. The humidity difference is [+2%, +1%, +1%, +2%, +3%, +4%], which is continuously positive and the amplitude increases, so the direction is marked as "rising". Combine the humidity direction of each group with the cumulative drift value and match it with each layer number. For example, if the oxygen drift of layer L7 is -1.0% and the humidity direction is "rising", the corresponding humidity offset matching direction value for this layer is "negative - rising". If the oxygen drift of layer L10 is +0.9% and the humidity direction is "falling", its matching direction value is "positive - falling". The entire structure forms a key - value set of [(layer number, drift polarity, humidity direction)], which serves as the basic data for subsequent joint index judgment.

[0038] S403: Call the humidity offset matching direction value, calculate the joint change intensity index of the oxygen drift amplitude and the humidity direction, generate the opening and closing state codes for the corresponding layers according to the set priority judgment threshold, and obtain the nitrogen injection valve hierarchical opening and closing instructions; The specific calculation formula for the joint change intensity index of the oxygen drift amplitude and the humidity direction is as follows: ; Among them, represents the joint change intensity index of the oxygen drift amplitude and the humidity direction, represents the oxygen concentration drift amplitude of the j - th measurement, represents the humidity value of the j - th measurement, represents the humidity direction value of the j - th measurement, represents the reference humidity direction value of the j - th measurement, represents the total number of measurements; To calculate the joint change intensity index of the oxygen drift amplitude and the humidity direction, , first, monitor the changes in oxygen concentration and humidity direction in consecutive m measurements. Set the oxygen concentration drift for each monitoring as and the current humidity direction value , as well as the reference humidity direction value . The measurement data is obtained through precise environmental sensors to ensure the real-time and accuracy of the data.

[0039] : represents the amplitude of oxygen concentration drift in the jth measurement. For example, in a specific monitoring, it may be recorded that the oxygen concentration changes from 20.5% to 20.7%, then is 0.2%.

[0040] : represents the humidity value at the jth measurement, such as the humidity sensor reading is 35%.

[0041] and : is the current measured humidity direction value, while is the set reference humidity direction value. For example, the current measured humidity direction is north and the reference direction is south.

[0042] Suppose the data obtained in three consecutive measurements are: The 1st time: , , (north), (south); The 2nd time: , , (east), (west); The 3rd time: , , (west), (east); The calculation formula steps are as follows to calculate each item each time: The 1st time: ; The 2nd time: ; The 3rd time: ; Accumulate and take the absolute value: ; Finally, the combined change intensity index is 0.2315. This result shows that in continuous measurements, the oxygen drift amplitude and the change in humidity direction have a significant combined effect. This indicator is used to determine whether the operating status of the equipment needs to be adjusted according to environmental changes, such as opening and closing the nitrogen injection valve to ensure stable operation of the system. In this way, the system responds to environmental changes and optimizes operation.

[0043] The specific steps of S5 are:

[0044] S501: based on the coordinates of the action area of ​​the nitrogen injection valve layered opening and closing instructions, synchronously obtain the real-time monitoring value of the inclination sensor in the corresponding space, extract the offset data in the axial dimension, and classify them according to the layer number to obtain the axial inclination offset; First, read the layered position coordinate information of the executed opening and closing action, extract each layer number with the opening and closing status of "open" or "closed" and its position coordinate point (x, y, z) in three-dimensional space, for example, the corresponding coordinates of L4 layer are (2.5, 3.0, 1.2), and those of L7 layer are (2.5, 3.0, 2.1). While confirming these coordinates, synchronously obtain real-time inclination data from the inclination sensor installed in the same space. The monitoring value returned by the inclination sensor is the angle offset data in the X-axis and Y-axis directions, in degrees, and the sampling frequency is set to once every 10 seconds. The extracted data is matched and stored according to each coordinate point. If the inclination angle obtained by L4 layer within the sampling period is X The X-axis is 0.8° and the Y-axis is 0.5°, then the data is bound to its coordinates (2.5, 3.0, 1.2), and then the axial offset values ​​of the inclination values ​​need to be unified. The square root of the offset amplitudes of the X-axis and the Y-axis is summed up and converted into a synthetic axial offset. That is, the synthetic axial inclination offset in the above example is √(0.8²+0.5²)=0.943°. This process is performed on all layered coordinate points in sequence and grouped according to the layer number. The inclination offsets of layers L4 to L12 are integrated into a structural map, such as L4→0.943°, L5→1.012°, and L6→0.765°, forming an axial inclination offset data set with layer numbers and corresponding synthetic inclination values.

[0045] S502: According to the axial inclination offset, the wind speed monitoring data of the air supply duct in the area is matched, the inclination value and the wind speed value consistent with the coordinates of the diversion dead angle area are extracted, and the two are multiplied according to the coordinate pair to obtain the airflow influence of the diversion dead angle; According to the axial inclination offset obtained in the previous step, for each coordinate position, match the sampling values of the wind speed sensors of the air supply ducts deployed in this area. The wind speed sensors collect the wind speed once per minute, with the unit of meters per second, and align it with the sampling time point of the inclination sensor to ensure data synchronization. For example, at the coordinate (2.5, 3.0, 1.5) on the L5 layer, the wind speed is 1.6 m / s, and the corresponding inclination offset is 1.012°. Then, the product of these two values needs to be calculated, that is, the airflow influence amount of the dead corner is 1.012×1.6 = 1.619, with the unit of (°·m / s). This product value represents the intensity of the superposition effect of the airflow at this position affected by the inclination. Perform the same operation for each layer, form paired data from the inclination offset value and the wind speed value at each coordinate position and calculate the product. After all coordinate points are processed, a mapping table of the airflow influence amount is formed, and the table structure is [(x, y, z), airflow influence value]. Example data are [(2.5, 3.0, 1.2), 1.320], [(2.5, 3.0, 1.5), 1.619], [(2.5, 3.0, 1.8), 1.448]. If the inclination offset at a certain position exceeds 3.0° and the wind speed is less than 1.0 m / s, then the product value is less than 1.0, and it is determined to be in the critical state of the dead corner. Mark such values for subsequent implementation of the ventilation adjustment rule. Set the critical threshold to 1.2, which is derived from the phenomenon of core ventilation obstruction caused by the combination of low wind speed and sudden inclination mutation at the failure boundary point of the airflow channel in multiple batches of storage structures. After retaining three decimal places for all product results, a complete set of airflow influence amounts of the diversion dead corner is formed.

[0046] S503: Invoke the airflow influence amount of the diversion dead corner, combine the coordinate distribution range of the diversion points, assemble the wind speed product values by region and generate the ventilation adjustment instruction structure corresponding to the position, establish a numerical configuration set for regulating the flow channel, and obtain the automated controlled atmosphere scheme for the cigarette stack; After calling the data of the air flow influence amount in the dead corner of the diversion, according to the spatial coordinate distribution of each diversion point, the air flow influence amount values belonging to the same area are aggregated to form an area air flow distribution group. The aggregation rule delimits the area boundary in the xy plane, and each 0.5 m is divided into a unit area. All levels in the z-axis direction are combined and statistically analyzed one by one. For example, within the range of area R1, there are three points z = 1.2, 1.5, and 1.8, and their corresponding air flow influence values are 1.320, 1.619, and 1.448. Then, a wind speed product value array [1.320, 1.619, 1.448] is constructed for this area, and the average value of this array is calculated to be 1.462. If this value is less than the set lower adjustment threshold of 1.5, the air supply speed of this area needs to be increased. If it is greater than the upper threshold of 2.5, the air supply speed needs to be decreased. The adjustment strategy is set to adjust the wind speed by 5% for every 0.2 unit deviation. That is, if the average value is 1.3, it is 0.2 different from 1.5, and the wind speed needs to be increased by 5%. If the average value is 1.0, the difference is 0.5, and the adjustment ratio is 12.5%, rounded up to 15%. Finally, an adjustment instruction structure [R1, +15%] is generated for area R1. All areas are calculated and allocated the adjustment intensity according to this method, and the results are summarized into a mapping table of area numbers and adjustment ratios, such as R1 → +15%, R2 → 0%, R3 → -10%. This adjustment ratio will be mapped into control instructions and sent to each fan sub-module in the format of [(x, y), adjustment ratio]. After the structure output, it is a numerical configuration set for the overall air flow regulation of the tobacco stack, and finally an automated controlled atmosphere scheme for the tobacco stack is obtained.

[0047] Please refer to Figure 2 , an automated controlled atmosphere system for a three-dimensional tobacco stack, comprising: The spectrum recognition module collects the gas fluctuation spectrum amplitude in the frequency band associated with the metabolic activity of mildew microorganisms in the sealed cavity of the tobacco stack, compares the amplitude differences in the same frequency band within adjacent sampling periods, arranges the difference values according to the frequency band serial numbers to form a continuous gradient set, and screens the number of frequency bands with amplitude changes to obtain the mildew activity gradient identifier of the tobacco stack; The heat warning module, based on the mildew activity gradient identifier of the tobacco stack, extracts the gradient fluctuation value of the surface heat release frequency band, counts the number of times of continuously deviating from the stored safe frequency domain threshold, and determines whether the number of deviations reaches the preset trigger condition to obtain the surface fermentation heat warning instruction; The wind speed instruction module calls the surface fermentation heat warning instruction, extracts the real-time monitoring value of the differential stratified oxygen concentration of the tobacco stack, and combines the downward trend of the mildew activity gradient identifier to match the spatial coordinates of the corresponding stacking density stratification to obtain the stratified rotation speed instruction of the circulation fan; The nitrogen injection control module, according to the feedback signal of the stratified rotation speed instruction of the circulation fan, detects the cumulative offset value of the drift amount of the corresponding stratified oxygen sensor, extracts the humidity distribution gradient direction, and calculates the cross-change intensity of the two to obtain the stratified opening and closing instruction of the nitrogen injection valve; Based on the action area of the hierarchical opening and closing instruction of the nitrogen injection valve, the scheme generation module obtains the axial offset of the regional accumulation inclination monitoring value, extracts the wind speed value of the air supply pipeline, calculates the product effect of the inclination offset and the wind speed, and obtains the automated controlled atmosphere scheme for the tobacco stack.

[0048] The above are only the preferred embodiments of the present invention, and the present invention is not limited to other forms. Any person skilled in the art may use the disclosed technical content to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as the technical solution content of the present invention is not departed from, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. An automated gas conditioning method for a three-dimensional cigarette stack, characterized in that: The following steps are involved: S1: In the sealed cavity of the cigarette stack, the amplitude changes of the frequency bands associated with the metabolism of moldy microorganisms in the gas spectrum are collected, the amplitude differences of the same frequency bands in adjacent sampling periods are compared, and the frequency bands are arranged into a gradient feature set according to the sequence number to generate a gradient mark of the cigarette stack moldy activity; S2: Based on the tobacco stack mildew activity gradient marker, locate the gradient fluctuation range of the tobacco fermentation heat release frequency band, calculate the number of consecutive deviations between the frequency band gradient value and the safety frequency domain threshold, and generate a surface fermentation heat warning instruction; S3: calling the surface fermentation heat warning instruction, extracting the layered monitoring value of the oxygen concentration in the cigarette stack, combining the attenuation trend of the mildew activity gradient, matching the layered topological structure of the stacking density, and generating the layered speed instruction of the circulating fan; S4: According to the operation feedback of the circulating fan stratified speed command, the accumulated offset value of the oxygen sensor stratified drift is detected, and the compensation priority is calculated in combination with the humidity stratified gradient change to generate the nitrogen injection valve stratified opening and closing command; S5: Based on the action coordinates of the stratified opening and closing instructions of the nitrogen injection valve, the axial offset of the stacking inclination angle is synchronously obtained, the product effect index with the stratified air supply wind speed is calculated, and an automated gas conditioning plan for the smoke stack is generated.

2. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 1, characterized in that: The mildew activity gradient identifier of the cigarette stack is specifically the amplitude sequence of the associated frequency band and the mildew activity difference gradient; the surface fermentation heat warning instruction includes the deviation frequency threshold, the safety threshold interval, and the warning level parameter; the circulating fan stratified speed instruction specifically refers to the stratified oxygen concentration distribution value, the mildew activity attenuation gradient, and the density stratified topological coordinates; the nitrogen injection valve stratified opening and closing instruction includes the oxygen sensor drift accumulation, the humidity distribution gradient vector, and the compensation priority sequence; the automated gas conditioning scheme of the cigarette stack is specifically the stacking inclination offset parameter, the diversion dead angle coordinate mapping, and the stratified wind speed control coefficient.

3. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 1, characterized in that: The specific steps of S1 are: S101: Collect gas fluctuation spectrum data of continuous time periods in the sealed cavity of the cigarette stack, extract characteristic frequency bands associated with the metabolism of moldy microorganisms, extract amplitude parameters corresponding to the frequency band numbers, establish a corresponding relationship between the frequency band numbers and the amplitudes, and obtain a frequency band amplitude matrix; S102: calling the amplitude sequence of the same frequency band in adjacent sampling periods in the frequency band amplitude matrix, calculating the absolute difference of the amplitudes of adjacent periods in the corresponding frequency band, arranging the difference values ​​according to the frequency band sequence number, and obtaining a frequency band difference sequence set; S103: According to the difference values ​​in the frequency band difference sequence set, a gradient change array is constructed according to the frequency band number, the quantitative situation characterizing the mildew activity is identified, and a gradient identifier of the mildew activity of the cigarette stack is obtained.

4. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 1, characterized in that: The specific steps of S2 are: S201: based on the tobacco pile mildew activity gradient mark, extract the frequency number in the frequency band corresponding to the surface tobacco, compare the numerical sequence corresponding to the number, analyze the continuous interval range of the fermentation heat release related frequency band in the gradient mark, and obtain the surface frequency band gradient interval; S202: calling the gradient value corresponding to the frequency band in the surface frequency band gradient interval, comparing the offset direction between the frequency band value and the threshold value based on the stored safe frequency domain threshold value as a judgment reference, counting the number of offsets occurring in continuous cycles, and obtaining the number of continuous offsets of the frequency band; S203: judging whether the set trigger condition is met according to the number of consecutive frequency band offsets, and if so, generating a warning status code according to the three-level warning rule, and constructing a command output based on the current cycle to obtain a surface fermentation heat warning command.

5. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 1, characterized in that: The specific steps of S3 are: S301: calling the surface fermentation heat warning instruction, extracting the oxygen concentration sensor monitoring value corresponding to the layer position in the cigarette stack, classifying and sorting all sensor data according to the collection time and layer height, establishing a mapping data table between oxygen concentration and layer level, and obtaining the oxygen concentration layer distribution value; S302: extracting the mildew activity gradient data sequence in the corresponding time period of each layer according to the oxygen concentration layer distribution value, calculating the sequence difference, and determining the decreasing amplitude change trend in the continuous time period, quantifying the gradient difference trend of each layer into a percentage form, and obtaining the gradient decreasing rate trend value; S303: combining the gradient descent rate trend value and the oxygen concentration stratified distribution value, matching the spatial coordinate system in the tobacco stacking density topological structure, calculating the ventilation control error characteristic value at each coordinate level, and obtaining the circulating fan stratified speed instruction.

6. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 5, characterized in that: The calculation formula of the ventilation control error characteristic value at each coordinate level is specifically: ； in, represents the ventilation control error eigenvalue in the spatial coordinates of the i-th row, j-th column and k-th layer, Represents the weighted average of the gradient descent rate trend value at the coordinate point in the i-th row and j-th column, Represents the standard deviation of the oxygen concentration distribution value at the i-th row and j-th column coordinate point. represents the tobacco packing density value corresponding to the spatial position of the i-th row, j-th column and k-th layer, Represents the gradient of the flow path change in the coordinate area of ​​the i-th row and j-th column, Represents the initial fan control template value corresponding to the coordinate of the i-th row, j-th column, and k-th layer.

7. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 1, characterized in that: The specific steps of S4 are: S401: extracting the monitoring value of the stratified oxygen sensor according to the operation feedback signal of the stratified speed command of the circulating fan, calculating the difference in the continuous cycle according to the time series, and performing cumulative difference statistics with the initial reference value to obtain the cumulative amplitude of oxygen drift; S402: extracting the layered humidity monitoring value in the stack according to the accumulated amplitude of oxygen drift, calculating the humidity difference between the layers, and determining the gradient change direction, corresponding to the drift amplitude according to the layer number, and obtaining the humidity offset matching direction value; S403: calling the humidity offset matching direction value, calculating the joint change intensity index of the oxygen drift amplitude and the humidity direction, generating the corresponding layered opening and closing state code according to the set priority judgment threshold, and obtaining the nitrogen injection valve layered opening and closing instructions.

8. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 7, characterized in that: The calculation formula of the combined change intensity index of the oxygen drift amplitude and humidity direction is specifically: ； in, An index representing the combined change intensity of oxygen drift amplitude and humidity direction, represents the drift amplitude of oxygen concentration measured for the jth time, represents the humidity value measured for the jth time, represents the humidity direction value measured for the jth time, represents the reference humidity direction value of the jth measurement, Represents the total number of measurements.

9. The automatic gas conditioning method for a three-dimensional cigarette stack according to claim 1, characterized in that: The specific steps of S5 are: S501: based on the coordinates of the action area of ​​the nitrogen injection valve layered opening and closing instructions, synchronously obtain the real-time monitoring value of the inclination sensor in the corresponding space, extract the offset data in the axial dimension, and collect them according to the layer number to obtain the axial inclination offset; S502: According to the axial inclination angle offset, the wind speed monitoring data of the air supply duct in the area is matched, the inclination angle value and the wind speed value consistent with the coordinates of the diversion dead angle area are extracted, and the two are multiplied according to the coordinate pair to obtain the airflow influence of the diversion dead angle; S503: Call the airflow influence amount of the diversion dead corner, combine the coordinate distribution range of the diversion point, assemble the wind speed product value by area and generate the ventilation adjustment instruction structure of the corresponding position, establish a numerical configuration set for regulating the circulation channel, and obtain the automatic air conditioning plan of the chimney.

10. An automatic gas conditioning system for a three-dimensional cigarette stack, characterized in that: According to an automated gas conditioning method for a three-dimensional cigarette stack according to any one of claims 1 to 9, the system comprises: The spectrum recognition module collects the gas fluctuation spectrum amplitude of the frequency band associated with the metabolism of moldy microorganisms in the sealed cavity of the cigarette stack, compares the amplitude difference of the same frequency band in adjacent sampling periods, arranges the difference values ​​according to the frequency band sequence number, forms a continuous gradient set, screens the number of amplitude change frequency bands, and obtains the cigarette stack mold activity gradient identification; The fever warning module extracts the gradient fluctuation value of the surface heat release frequency band based on the mildew activity gradient mark of the cigarette stack, counts the number of consecutive deviations from the stored safety frequency domain threshold, determines whether the number of deviations reaches the preset trigger condition, and obtains the surface fermentation heat warning instruction; The wind speed instruction module calls the surface fermentation heat warning instruction, extracts the real-time monitoring value of the oxygen concentration of the differentiated layers of the cigarette stack, combines the downward trend of the mold activity gradient mark, matches the spatial coordinates of the corresponding stacking density layer, and obtains the circulating fan layer speed instruction; The nitrogen injection control module detects the accumulated offset value of the drift of the corresponding stratified oxygen sensor according to the feedback signal of the stratified speed command of the circulating fan, extracts the direction of the humidity distribution gradient, calculates the cross-change intensity of the two, and obtains the stratified opening and closing command of the nitrogen injection valve; The solution generation module obtains the axial offset of the regional stacking inclination monitoring value based on the action area of ​​the nitrogen injection valve stratified opening and closing instructions, extracts the wind speed value of the air supply duct, calculates the product effect of the inclination offset and the wind speed, and obtains the automated gas conditioning solution for the chimney.

Images