Medicinal material drying method and system based on big data

Abstract

The invention relates to the technical field of big data processing, in particular to a medicinal material drying method and system based on big data, and the method comprises the following steps: obtaining a sample moisture value to construct a time sequence, calculating rate difference classification to generate a number, extracting a trend section matching configuration table, and collecting environmental parameters to construct an offset sequence. And combining gradient concentration and deformation generation residual error to judge abnormity, extracting an offset trend, grouping, and outputting a drying scheme. According to the method, a classification system is constructed by extracting a sample moisture rate difference value, initial configuration matching is formed by combining a trend section and an affiliation relationship, an offset sequence is constructed by fusing producing area climate parameters and drying key time points, and residual data is generated by linking a moisture gradient, a concentration ratio and a structural deformation angle. According to the method, trend deviation identification and adjustment grouping operation are completed, dynamic response to sample features and environment variables is realized, identification precision, adjustment accuracy and stable control capability in drying processing are improved, and drying adaptability and process decision efficiency under complex conditions are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of big data processing, and in particular to a medicinal material drying method and system based on big data. Background Art

[0002] The field of big data processing technology involves the entire process of efficiently collecting, storing, managing, analyzing, and applying massive and diverse data. The core issues in this technology include the cleaning, integration, modeling, analysis, and visualization of large-scale data, and are widely used in multiple industries such as healthcare, agricultural production, industrial manufacturing, and financial analysis. Its technical foundation relies on distributed storage technology, parallel computing architecture, and complex data mining and pattern recognition mechanisms, which can achieve unified management of structured and unstructured data and deep value extraction. Among them, traditional medicinal material drying methods refer to the process of artificially or mechanically removing the moisture content of Chinese medicinal materials after they are harvested in order to facilitate preservation and enhance their efficacy. The technical issues addressed are the large differences in medicinal material types during the drying process, the difficulty in unifying the drying curve, and the resulting unstable quality control. Traditional medicinal material drying technology usually sets the heating temperature and time based on experience, and uses hot air drying or sun drying methods to complete the dehydration process by controlling the air flow rate and temperature in the heating equipment.

[0003] Existing technologies mainly rely on manually set parameters, and have not established sample classification standards for differences in moisture distribution in the internal structure of Chinese medicinal materials. When processing medicinal materials from different environmental conditions, the spatial hierarchical characteristics of moisture migration are easily ignored. Faced with humidity, temperature fluctuations or circadian rhythm differences between production areas, there is a lack of an identification mechanism for dynamic nodes in the drying process, such as inflection points and termination moments, and it is impossible to establish an accurate mapping between offset timing and behavioral responses. In addition, there is a lack of a detection mechanism for the correlation between structural deformation and moisture ratio, resulting in processing lags and error accumulation in the middle and late stages of drying. For example, in an environment with unstable humidity, unified drying parameters cannot adapt to changes in internal and external moisture gradients in a timely manner, resulting in problems such as uneven sample quality and structural stress concentration, affecting the overall drying quality and stability. Summary of the Invention

[0004] The purpose of the present invention is to solve the shortcomings of the prior art and propose a medicinal material drying method based on big data.

[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solution: a medicinal material drying method based on big data, comprising the following steps: S1: Obtain the moisture values ​​of the surface and center of the Chinese medicinal material samples, construct a time series based on the sampling interval, calculate the moisture change rate at adjacent time points, extract the rate difference sequence between samples, perform a comparison to see if they exceed the judgment threshold, classify the samples pairwise, and obtain a difference group number set; S2: Divide the samples according to the difference group number set, extract the rate sequence to calculate the average trend, determine the segment to which the trend belongs, match the set number interval, combine the corresponding number and sample ownership information, and obtain the initial configuration matching table; S3: Extracting the sample origins from the initial configuration matching table, collecting humidity, temperature frequency, and day / night difference, obtaining the inflection point and end time in the drying curve, calculating the humidity and inflection point time difference, and the end time offset, constructing two sorted sequences, and performing index difference comparison to obtain an origin offset sequence set; S4: Extract the offset samples from the origin offset sequence set, call the moisture gradient map and concentration ratio sequence of the drying process, pair the structural deformation angle sequence, construct the residual sequence at the same time point, determine whether it exceeds the error limit continuously, record the abnormal number and time period, and obtain a list of deformation abnormal sections.

[0006] As a further solution of the present invention, the difference grouping number set includes the moisture change rate difference, the sample classification number, and the rate difference threshold identifier; the starting configuration matching table includes the grouping number correspondence, the sample trend segment number, and the configuration matching number; the origin offset sequence set includes the humidity time offset, the end time offset, and the origin sorting index difference; the deformation abnormality segment list includes the abnormal sample number, the abnormal time period number, and the residual over-limit mark.

[0007] As a further solution of the present invention, the step of obtaining S1 is: S101: Based on the time points at which the surface moisture value and the center moisture value of the Chinese medicinal material sample are obtained, the change in moisture value at adjacent time points is calculated, and the difference rate between the surface moisture value and the center moisture value at adjacent time points is determined. By constructing a time series, the trend of moisture difference change at consecutive time points is obtained, and a moisture difference change rate sequence is generated; S102: Calling the rate values ​​at the same time between any two samples in the moisture difference change rate sequence, constructing a time series difference degree based on the time point difference change, calculating and obtaining a rate difference index between samples, and establishing a sample rate difference index set; S103: Based on the difference index value between any two samples in the sample rate difference index set, call the set moisture change difference threshold, judge whether the difference value exceeds the threshold in turn, classify according to whether it exceeds, divide the samples that meet the same classification conditions into a unified number group, and obtain a difference group number set.

[0008] As a further solution of the present invention, the step of obtaining S2 is: S201: Divide the Chinese medicinal material samples based on the differential group number set, call the moisture change rate sequence corresponding to the sample under the number, perform a sum operation on the sample rate values ​​according to the time dimension and divide it by the number of samples, obtain the average rate trend value sequence of the time points under the differential grouping, and generate a group average trend sequence; S202: Calling the time series mean and series slope under multiple numbers in the grouped average trend sequence, calculating and obtaining the trend index value under multiple numbers, calling the set segment standard according to the position of the trend index value within the segment range, determining the segment number to which it belongs, and generating a trend segment number sequence; S203: According to the trend segment number sequence, each number is combined with the corresponding sample affiliation number and group number to establish a corresponding structure of sample affiliation, trend segment and group number to obtain an initial configuration matching table.

[0009] As a further solution of the present invention, the acquisition step of S3 is: S301: Obtain the sample origin in the initial configuration matching table, collect humidity, temperature frequency and day and night difference, extract the inflection point time and end time in the drying curve, calculate the humidity and inflection point time difference and make corrections to obtain the drying process climate offset value sequence; S302: calling the inflection point time difference and the end time offset value in the drying process climate offset value sequence, constructing two sorting sequences and extracting index numbers, calculating the difference between the indexes, and obtaining a sorting index offset sequence; S303: Rearranging the order of the samples according to the index offset value of each origin in the sorted index offset sequence in combination with the sample origin number information, establishing a corresponding structure sequence of samples and origins, and obtaining an origin offset sequence set.

[0010] As a further solution of the present invention, the acquisition step of S4 is: S401: extracting the offset samples from the origin offset sequence set, calling the corresponding drying process moisture gradient map and concentration ratio sequence, pairing the structural deformation angle sequence, performing an operation on the difference between the moisture gradient value and the structural angle value at the same time point, constructing a difference set at the time point, and obtaining a residual sequence at the same time point; S402: Calling the time period difference in the residual sequence at the same time point, comparing it with the structural error limit, screening the time periods that continuously exceed the limit, recording the corresponding sample numbers and continuous time period information, and obtaining the residual abnormal segment number set; S403: According to the sample numbers and corresponding time periods recorded in the residual abnormal segment number set, the offset sample identifiers and abnormal time periods are integrated, a corresponding structure between numbers and time periods is established, and a list of abnormal deformation sections is obtained.

[0011] As a further embodiment of the present invention, the method further comprises: S5: Calling samples from the abnormal deformation section list, extracting the phase moisture gradient direction change and angle change rate, calculating the offset trend value of the two, determining whether they enter the trend deviation range, grouping them into an adjustment result set, and obtaining a medicinal material drying treatment plan; The medicinal material drying treatment scheme includes a moisture gradient direction offset value, an angle change rate offset value, and a trend deviation identification result; The steps to obtain S5 are: S501: calling samples in the abnormal deformation section list, extracting the moisture gradient direction change values ​​and structural angle change rate values ​​of multiple stages, synchronously integrating the corresponding data according to time points, and establishing a stage gradient and angle sequence set; S502: Calculate the offset trend value between the moisture gradient direction change and the angle change rate based on the data of multiple time points in the stage gradient and angle sequence, determine whether it falls within the set trend deviation range, select the sample numbers that meet the conditions, and obtain the trend offset mark sequence; S503: calling the offset sample number in the trend offset mark sequence, grouping the corresponding samples and adjustment schemes, building a correspondence between the sample number and the processing structure, and obtaining the medicinal material drying processing scheme.

[0012] The medicinal material drying system based on big data includes: The moisture grouping module obtains the surface and center moisture values ​​of the sample, constructs a time series, calculates the rate of change, extracts the rate difference sequence between samples, determines whether it exceeds the set standard, classifies the sample pairs, and obtains a set of difference group numbers; The trend configuration module divides the samples based on the difference group number set, extracts the rate sequence, calculates the average trend, determines the segment to which it belongs, matches the set number, combines the attribution information, and obtains the starting configuration matching table; The origin offset module extracts the origin in the starting configuration matching table, collects humidity and temperature parameters, obtains the drying inflection point and end time, calculates and sorts the time difference, performs index difference comparison, and obtains an origin offset sequence set; The deformation identification module extracts offset samples based on the origin offset sequence set, calls the moisture gradient map and concentration ratio sequence, pairs the structural angle values, constructs a residual sequence, determines whether the error limit is continuously exceeded, records abnormal samples and time periods, and obtains a list of abnormal deformation sections; The adjustment scheme module calls the abnormal deformation section list, extracts the moisture gradient direction and angle change rate, calculates the offset trend, determines whether it enters the deviation range, groups the sample numbers, and obtains the medicinal material drying treatment plan.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, a classification system is constructed by extracting the difference in sample moisture rate, an initial configuration match is formed by combining trend segments and attribution relationships, an offset sequence is constructed by integrating the climate parameters of the origin and the key drying time points, and residual data is generated by linking the moisture gradient, concentration ratio and structural deformation angle. The trend offset identification and adjustment grouping operations are completed, and a dynamic response to sample characteristics and environmental variables is achieved. The recognition accuracy, adjustment accuracy and stable control ability in the drying process are improved, and the drying adaptability and process decision-making efficiency under complex conditions are enhanced. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flow chart of the main steps of the present invention; Figure 2 This is the S1 flow chart of the present invention; Figure 3 This is the flow chart of S2 of the present invention; Figure 4 This is the flow chart of S3 of the present invention; Figure 5 This is the flow chart of S4 of the present invention; Figure 6 This is the S5 flow chart of the present invention. DETAILED DESCRIPTION

[0015] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0016] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0017] See also Figure 1 , a medicinal material drying method based on big data, comprising the following steps: S1: Obtain the moisture values ​​of the surface and center of the Chinese medicinal material samples, construct a time series based on the sampling interval, calculate the moisture change rate at adjacent time points, extract the rate difference sequence between samples, perform a comparison to see if they exceed the judgment threshold, classify the samples pairwise, and obtain a difference group number set; S2: Divide the samples according to the difference group number set, extract the rate sequence to calculate the average trend, determine the segment to which the trend belongs, match the set number interval, combine the corresponding number and sample ownership information, and obtain the starting configuration matching table; S3: Extract the sample origins from the starting configuration matching table, collect humidity, temperature frequency, and day / night difference, obtain the inflection point and end time of the drying curve, calculate the humidity and inflection point time difference and end time offset, construct two sorted sequences, and perform index difference comparison to obtain the origin offset sequence set; S4: Extract the offset samples from the origin offset sequence set, call the moisture gradient map and concentration ratio sequence of the drying process, pair the structural deformation angle sequence, construct the residual sequence at the same time point, determine whether it exceeds the error limit continuously, record the abnormal number and time period, and obtain a list of deformation abnormal sections; S5: Call the samples in the list of abnormal deformation sections, extract the direction change and angle change rate of the moisture gradient, calculate the offset trend value of the two, determine whether they enter the trend deviation range, group them into an adjustment result set, and obtain the medicinal material drying treatment plan.

[0018] The difference grouping number set includes the moisture change rate difference, sample classification number, and rate difference threshold identifier; the starting configuration matching table includes the grouping number correspondence, the trend segment number to which the sample belongs, and the configuration matching number; the origin offset sequence set includes the humidity time offset, the end time offset, and the origin sorting index difference; the deformation abnormality segment list includes the abnormal sample number, the abnormal time period number, and the residual over-limit mark; the medicinal material drying treatment plan includes the moisture gradient direction offset value, the angle change rate offset value, and the trend deviation identification result.

[0019] See also Figure 2 , the steps to obtain S1 are: S101: Based on the time points at which the surface moisture value and the center moisture value of the Chinese medicinal material sample are obtained, the change in moisture value at adjacent time points is calculated, and the difference rate between the surface moisture value and the center moisture value at adjacent time points is determined. By constructing a time series, the trend of moisture difference change at consecutive time points is obtained, and a moisture difference change rate sequence is generated; Based on the time points for obtaining the surface moisture value and the center moisture value of the Chinese medicinal material samples, it is first necessary to clarify the sample detection cycle and detection method during actual sampling. For example, in the drying experiment of a batch of astragalus samples, it is set to collect moisture data every 15 minutes. The sampling period covers a total of 60 minutes, including 5 time points of 0, 15, 30, 45 and 60 minutes. At each time point, the surface moisture value of the sample is detected by a non-contact infrared moisture meter, and the moisture content at the center of the sample is measured by a penetrating capacitive sensor, and the following is obtained. Data: The surface moisture values ​​at the five time points were 41.3%, 38.2%, 34.9%, 31.0%, and 27.6%, respectively. The center moisture values ​​were 48.9%, 45.4%, 41.5%, 37.9%, and 34.0%, respectively. The moisture value change was calculated for the data between each two adjacent time points, that is, the moisture value at the next moment was subtracted from the moisture value at the previous moment. The surface change sequence was 3.1%, 3.3%, 3.9%, and 3.4%, and the center change sequence was 3.5%, 3.9%, and 3.6%. %, 3.9%, and then the difference analysis of the surface change value and the center change value in each time period was performed to obtain the difference value of the moisture change rate. In the first time period (0 to 15 minutes), the surface dropped to 3.1% and the center dropped to 3.5%, with a difference of 0.4% between the two; in the second time period, the surface changed to 3.3% and the center changed to 3.9%, with a difference of 0.6%; the difference in the third period was 3.6% in the center minus 3.9% in the surface, which was -0.3%; the difference in the fourth period was 3.9% in the center minus 3.4% in the surface, which was 0.5%. ; All difference values ​​are combined into a difference rate sequence of {0.4, 0.6, -0.3, 0.5}, and a time series trend graph is drawn for the sequence, with the horizontal axis as the time point segment (15 minutes, 30 minutes, 45 minutes, 60 minutes), and the vertical axis as the moisture rate difference value. In practical applications, this curve is used to show the degree of moisture transfer from the outside to the inside or from the inside to the outside during the dehydration process of Chinese medicinal materials. Its changing trend can be used as a direct basis for analyzing the drying uniformity of samples under different structural characteristics, thereby providing the original rate characteristic basis for the subsequent classification of similar samples.

[0020] S102: Call the rate value at the same time between any two samples in the moisture difference change rate sequence, and construct the time series difference based on the time point difference change, using the formula: ; Calculate and obtain the rate difference index between samples, and establish a sample rate difference index set; in, Representative samples With sample The rate difference indicator, and Represent samples With sample In time The rate of change of moisture, is the length of the time series, To avoid the fine-tuning constant introduced by the denominator being zero, is the adjustment factor of the sample normalized difference, The denominator balance constant for the normalized difference; After obtaining the moisture difference change rate series for multiple samples, the rate value at the same time between any two samples needs to be calculated to quantify the degree of rate difference between the samples. The formula is as follows: ; Assume that the moisture difference rate sequence of sample A is {0.6, 0.5, 0.8}, and that of sample B is {0.4, 0.9, 0.6}, set , the item-by-item calculations are as follows: The first part of the calculation: Item 1: ; Item 2: ; Item 3: ; Find the sum and average: ; The second part of the calculation: Sum of squared differences: ; The rate sum plus the constant: ; ; Normalize some results: ; Final indicator: ; The calculation shows that the difference index between samples A and B at the corresponding three time points is 0.3176, reflecting the difference in their moisture removal rate changes.

[0021] S103: Based on the difference index value between any two samples in the sample rate difference index set, call the set moisture change difference threshold, determine whether the difference value exceeds the threshold in turn, classify according to whether it exceeds, and divide the samples that meet the same classification conditions into a unified number group to obtain a difference group number set; After obtaining the difference in moisture change rate between each sample, in order to classify the samples, a numerical value needs to be set as a reference for difference judgment, that is, a threshold. In this embodiment, the threshold is set to 0.25 percentage points, and the difference values ​​between samples are judged one by one in a textual manner. For example, the difference in moisture change rate between sample A and sample B at three groups of time points is 0.32 after integration, which exceeds the set threshold. Therefore, A and B are considered to be quite different and should not be classified into the same group. They are numbered as category 1 and category 2 respectively. Then, sample C and sample D are compared. The difference in moisture rate is 0.08, which is lower than the threshold, indicating that their dehydration characteristics are similar and can be classified into the same category. The number is set to category 3, and the samples are further judged. The difference between sample E and sample F is 0.41, far exceeding the threshold, indicating that their drying behaviors are significantly different. They should be numbered as categories 4 and 5, respectively. This classification operation continuously compares the difference values ​​between samples and judges them against the set standards to form a set of grouping numbers based on samples. In practical applications, this can be used as a reference for formulating different drying strategies or parameter settings for specific groups in subsequent processes. The grouping number set is ultimately presented in the form of a one-dimensional array, such as {1, 2, 3, 3, 4, 5}. Each number represents a set of samples with similar moisture change behaviors. The classification operation uses text to decide whether each sample belongs to the same category, avoiding algorithm dependence and relying solely on the numerical comparison process to make decisions.

[0022] See also Figure 3 , the steps to obtain S2 are: S201: Divide the Chinese medicinal material samples based on the differential group number set, call the moisture change rate sequence corresponding to the sample under the number, perform a sum operation on the sample rate values ​​according to the time dimension and divide it by the number of samples, obtain the average rate trend value sequence of the time point under the differential group, and generate the group average trend sequence; First, we need to clarify the basic structure and grouping criteria of Chinese herbal medicine samples. For example, during the drying process of a certain Chinese herbal medicine, it can be divided into three groups: A1, A2, and A3, based on the origin, initial moisture content, or drying method. Each group contains five samples: S01 to S05, S06 to S10, and S11 to S15. Each sample corresponds to a sequence of moisture change rates recorded in minutes. Next, we call the rate sequence of these samples. For example, the rate of sample S01 over 10 minutes is: [0.25, 0.27, 0.29, 0.30, 0.28, 0.26, 0.24, 0.22, 0.20, 0.18], with the unit of percentage change rate per minute. Other samples such as S02 to S05 also have rate sequences of the same dimension. For each group number, such as group A1, a time point accumulation operation is performed. That is, for the first minute, the rate values ​​of S01 to S05 in the first minute are extracted, which are 0.25, 0.26, 0.24, 0.27, and 0.25, respectively. The total rate value is 1.27 after accumulation. Then, the average rate value of this time point is calculated as 0.254 by dividing the sample number 5. This operation is repeated at each time point, and the rate mean of all samples at the same time is counted point by point to complete the average rate calculation of 10 time points, finally forming a 10-dimensional average rate trend sequence for this group, which is recorded as: [0.254, 0.268, 0.282, 0.292, 0.280, 0.264, 0.248, 0.230, 0.212, 0.196]. This process is performed separately for each group. For example, for group A2, if the first-minute rates corresponding to samples S06 through S10 are 0.30, 0.31, 0.29, 0.32, and 0.30, the cumulative value is 1.52, with an average value of 0.304. Similarly, the sample rate values ​​at each time point are accumulated and averaged to obtain the 10-dimensional rate trend for A2. The same process is applied to group A3, forming the average trend of the sample rates for each group along the time dimension, thus forming the final group trend sequence data structure. At this point, all groups have completed the transformation from the original sample rate series to the time-based average summary, forming the original basic series that can be further used in trend indicator calculations, with clear group and time positioning attributes.

[0023] S202: Call the time series mean and series slope under multiple numbers in the grouped average trend sequence, using the formula: ; Calculate and obtain the trend indicator value under multiple numbers. According to the position of the trend indicator value within the segment range, call the set segment standard, determine the segment number to which it belongs, and generate the trend segment number sequence. in, Representative Group The trend indicator value of Representative Group In time The average rate value, is the mean of the average rate of the group at all times, is the time series length of the group, and are the control factors for trend shift and volatility respectively, To avoid division by zero constants; Call multiple numbers such as A1, A2, and A3 in the above grouped average trend sequence to obtain their time series mean and series slope, which are recorded as and Fitting trend, according to the formula: ; Calculate the trend indicator value, where is the trend indicator value of group k, is the average rate value of the kth group at time t, is the mean of the rate sequence of group k, is the length of time, , , is a constant.

[0024] Taking group A1 as an example, its average trend sequence is: ; Step 1: Calculation , which is the average value: ; Step 2: Calculate the absolute value and offset terms one by one: Take the first three items as an example: , ; , ; , ; And so on, the absolute values ​​of all 10 items and the squared offset items are calculated: Step 3: Calculate the standard deviation term: The sum of the squares of all offsets is approximately , the average square root of which is ; Step 4: Calculate the normalized mean term: ; Step 5: Substitute the formula for weighted calculation: ; Substitute each term and add them up one by one. The first three terms in this example are: Item 1: ; Item 2: ; Item 3: ; The sum of the 10 items is approximately 3.99, and finally: ; Step 6: Trend segment judgment: If the segment division rule is: 0.00–0.30 is low (number 1), 0.30–0.45 is medium (number 2), and 0.45 and above is high (number 3), then the current Belongs to the middle section, corresponding to number 2.

[0025] The results show that the overall change in moisture rate of this group is in an intermediate trend state, which can be used as an important indicator for subsequent group attribution mapping.

[0026] S203: According to the trend segment number sequence, each number is combined with the corresponding sample affiliation number and group number to establish a correspondence structure among sample affiliation, trend segment, and group number, thereby obtaining an initial configuration matching table; First, the group numbers that appear in each trend number sequence are back-mapped. For example, the group with trend segment number 2 is A1, the group with number 1 is A2, and the group with number 3 is A3. Then, for each group A1, A2, and A3, their corresponding sample number sets are searched respectively. For example, the sample numbers corresponding to A1 are S01, S02, and S03, the samples corresponding to A2 are S04, S05, and S06, and the samples corresponding to A3 are S07, S08, and S09. For each sample number, the group A number and the trend segment number corresponding to the group are recorded, and a clear correspondence structure between the three is constructed to generate a sample attribution table. The table entry contains three fields in structure, namely "sample number", "group number", and "trend segment number". The specific construction process traverses all sample numbers in turn and extracts the group number from the group number table to which it belongs. Then search for the trend number corresponding to the group from the trend segment sequence, and record the triplet after completing a match. For example, sample S01 matches (S01, A1, 2), S05 matches (S05, A2, 1), S08 matches (S08, A3, 3), and so on, until all samples are assigned and numbered. Finally, a fixed structure is formed with direct accessibility and consistency, which facilitates the subsequent rapid search for the trend segment and group number according to the sample number. Moreover, the matching table serves as the basic support structure for various classification, screening, and segmented identification in the system. It is independently formed without relying on the construction of additional data structures, and has high stability and consistency. In practical applications, it can be directly used for drying strategy allocation or sample screening basis, and supports subsequent incremental updates. Whenever a new sample number or trend number is added, an append operation can be performed in the structure to complete real-time maintenance.

[0027] See also Figure 4 , the steps to obtain S3 are: S301: Obtain the sample origin in the starting configuration matching table, collect humidity, temperature frequency and day and night difference, extract the inflection point time and end time in the drying curve, calculate the humidity and inflection point time difference and make corrections to obtain the drying process climate offset value sequence; Obtain the sample origin in the starting configuration matching table. First, perform field parsing on the configuration table, identify and extract the sample number and origin number in each sample entry. The origin number should be a unique identifier in the system, such as "A01" and "A02" to avoid confusion. After obtaining the sample origin, extract the humidity and temperature monitoring data of the origin in the historical period. The frequency of humidity and temperature collection is statistically analyzed as follows: retrieve the number of records on the same day from the dataset by hourly dimension. For example, humidity is recorded 4 times a day, which is accumulated to 40 times. If the data of a certain day is missing in the middle, it will not be counted to maintain the validity of the total frequency. The extraction of the day and night temperature difference is calculated by the difference between the daily maximum and minimum temperatures. It is smoothed using a 10-day sliding mean and finally the mean of the day and night temperature difference is obtained as one of the sample input parameters. In the process of constructing the drying curve, the trend line graph of humidity over time is used. According to the inflection point judgment rule, the time node with significant slope change is determined as the inflection point time point. For example, in the process of humidity dropping from 18.5g / kg to 11.3g / kg, if the unit hourly humidity decrease rate is in the first In the 6th hour, the humidity dropped from 1.0g / kg in the previous hour to 0.4g / kg, with a rate of change of 0.6. The threshold for judging the inflection point is set to 0.5, that is, if the rate of change is greater than 0.5, it is considered an inflection point. The threshold is based on the average standard deviation of the decline rate of the first 50 data in the statistical sample group, and is set with an upward float of 10%. The judgment standard for the end time of the drying process is that the subsequent humidity change fluctuates less than 0.01g / kg for 3 consecutive hours. For example, the humidity in the 14th, 15th and 16th hours is 10.4, 10.41 and 10.41 respectively. .42, the maximum difference between the three values ​​is 0.02, and the average change is 0.0067, which is lower than the set threshold of 0.01. Therefore, the 16th hour is considered to be the end time of drying. After obtaining the humidity difference, the difference between the initial humidity and the inflection point humidity is recorded as 7.2g / kg, and the inflection point time difference is calculated to be 6 hours. The average time statistics of the samples from the origin are 5 hours, and the correction coefficient is set to 5÷6=0.833. This value is used to correct the humidity difference result. The final recorded offset value is 6.0, forming a drying process climate offset value sequence containing multiple samples.

[0028] S302: Calling the inflection point time difference and the end time offset value in the drying process climate offset value sequence, constructing two sorting sequences and extracting index numbers, calculating the difference between the indexes, and obtaining a sorting index offset sequence; Call the inflection point time difference and end time offset value in the drying process climate offset value sequence. First, read the inflection point time difference and end time offset value of each sample from the sequence. The inflection point time difference is the difference between the inflection point occurrence time and the average inflection point time of all samples. The end time offset value is the difference between the sample end time and the standard end time (set to 10 hours). For example, if the inflection point time of a sample is 7 hours and the average inflection point time of all samples is 6 hours, the offset is +1 hour. If the end time is 14 hours, the offset is +4 hours. Construct the two offsets into two one-dimensional arrays respectively, and sort them in ascending order respectively. The sorting operation records the original sample number index corresponding to each value. After the sorting is completed, the two index lists are compared. For example, the inflection point sequence sorting The index is [3, 1, 2, 0], and the end time series is [1, 0, 2, 3]. The displacement difference of each sample index in the two series is calculated: the first sample is 3-1=2, the second sample is 1-0=1, the third sample is 2-2=0, and the fourth sample is 0-3=-3, obtaining the sorting index offset sequence [2, 1, 0, -3]. This difference is the basic data reflecting the degree of sorting offset, which is further used for subsequent sample structure adjustment. By comparing the offsets of multiple groups of samples, the offset judgment interval is set to [-5, 5]. That is, when the sample sorting offset is greater than 5 or less than -5, it is marked as a strong offset sample. The offset interval is set based on the sample group average number of samples being 20, and 5 is a 25% sorting change amplitude, which is within the acceptable range of variation.

[0029] S303: Rearranging the samples in order based on the index offset value of each origin in the sorted index offset sequence and the origin number information of the samples, establishing a corresponding structure sequence between the samples and the origins, and obtaining an origin offset sequence set; According to the index offset value of each origin sample in the sorting index offset sequence, the sample origin number information is combined to re-arrange the order. First, read the sorting index offset sequence and the corresponding table of sample number and origin number. The execution process is to scan the index offset value of each sample and record its corresponding origin number information. Based on the current index offset value, adjust the position of the sample in the sequence. For example, sample S1001 comes from origin A01, and the index offset value is +2, then it will be moved two places behind the original sequence in the new sequence. If the offset is a negative value, it will move forward the corresponding position. During the adjustment process, it is necessary to avoid repeated coverage or crossing Because of the error in the boundary, an intermediate array is used to temporarily store the adjustment process data, and finally a stable structural order is regenerated. All samples are numbered in sequence according to the rearranged order to form the corresponding structural sequence of samples and origins, and the origin offset sequence set is obtained. On this basis, the origins with significant displacement can be marked. For example, if the absolute value of the offset value of a sample from a certain origin is greater than 5, it is recorded as a high-displacement origin. The threshold of 5 is set as twice the average value calculated according to the standard deviation of the index offset values ​​in all samples to ensure that the mark is representative and actually different, and finally a sample-origin-offset value ternary structure is formed for subsequent classification processing or analysis.

[0030] See also Figure 5 , the steps to obtain S4 are: S401: Extract the offset samples from the origin offset sequence set, call the corresponding drying process moisture gradient map and concentration ratio sequence, pair the structural deformation angle sequence, perform calculations on the difference between the moisture gradient value and the structural angle value at the same time point, construct a difference set at the time point, and obtain the residual sequence at the same time point; When extracting the offset samples from the origin offset sequence set, the monitoring data of each origin is first read in sequence according to the regional number, and the records with significant structural displacement changes are identified. The maximum displacement within a 24-hour period is used as the criterion. If the displacement increment of a sample exceeds 3.0 mm between any two consecutive time points, it is marked as an offset sample. The offset sample number, time point information and displacement value are recorded in the offset sequence table. For example, for the sample with origin number P24, the displacement increases from 4.6 mm to 8.2 mm from the 10th hour to the 12th hour, with an increment of 3.6 mm, which meets the offset criterion. Therefore, the time point record is extracted as the offset sample. Then the moisture gradient map during the drying process is called. The map divides the moisture content values ​​at different depths according to the structural profile. For example, at the 12th hour of the P24 sample, the moisture content of the surface layer is 21.3%, the middle layer (5 cm deep) is 15.6%, and the core layer (10 cm deep) is 12.1%. The moisture content is obtained layer by layer at intervals of 1 cm to form a gradient curve. The moisture gradient from the surface to the core layer is calculated to be (2 1.3%-12.1%) / 10cm=0.92% / cm, and the concentration ratio sequence is formed by the outer-inner layer moisture content ratio, which is 21.3 / 12.1≈1.76 here. This ratio is matched with the subsequent deformation angle. The angle sequence records the structural deformation angle of the sample in each direction in three-dimensional space, for example, 2.3° in the horizontal direction, 2.0° in the vertical direction, and 1.2° in the thickness direction. Time synchronization is achieved by aligning the sensor timestamps, and then the moisture gradient value and the deformation angle value are compared at each time point. After normalization, the difference is calculated to obtain the residual value. For example, the gradient of 0.92% / cm is normalized to 0.65, and the angle mean (2.3+2.0+1.2) / 3=1.83° is normalized to 0.47, so the difference is 0.18. The same processing is performed at each time point to form a sample residual sequence such as {0.15, 0.18, 0.21, 0.22, 0.19} for subsequent analysis. Each data record is constructed based on the same time base to ensure a one-to-one correspondence between gradients and angles without mismatching.

[0031] S402: Call the time period difference in the residual sequence at the same time point, compare it with the structural error limit, filter out the time periods that continuously exceed the limit, record the corresponding sample number and continuous time period information, and obtain the residual abnormal segment number set; After obtaining the residual sequence, it is necessary to further identify the continuous time period differences that may indicate structural abnormalities. The structural error limit value is set to 0.17. This value comes from the statistical results of 300 groups of experiments in the early stage. During the drying process under the test conditions of temperature 55℃ and drying time 24 hours, when the residual continuously exceeds 0.17, more than 87% of the samples were detected with microcracks or permanent deformation in subsequent scans. Therefore, 0.17 is set as the judgment limit. When executing the screening process, the residual sequence is traversed according to the sliding window method, with each three consecutive time points as a window to detect whether they all exceed the limit value. For example, the residual sequence of sample P24 is {0.15, 0.18, 0.21, 0.22, 0.19}, and the 2nd, 3rd, and 4th items are {0 .18, 0.21, 0.22}, all greater than 0.17, and the starting time is recorded as the 11th hour, with a continuous length of 3 hours. In the judgment process, no average or inclusion logic is used, but a separate comparison operation is performed on each value. If the difference is greater than the limit, it is abnormal, such as 0.21>0.17 is marked as "abnormal", otherwise it is "normal". Only when three consecutive items are abnormal is the time period confirmed to be valid. All time periods that meet the conditions and their corresponding sample numbers are combined and recorded to form a residual abnormal segment number set. The recording format is such as (P24, 11~13). There can be multiple abnormal time periods in a sequence, and each segment is processed independently. The setting process of the error limit value has been clarified, which meets the experimental verification requirements and is repeatable and reasonable.

[0032] S403: Based on the sample numbers and corresponding time periods recorded in the residual abnormal segment number set, the offset sample identifiers and abnormal time periods are integrated to establish a corresponding structure between the numbers and time periods, thereby obtaining a list of abnormal deformation sections. When integrating the sample numbers and time periods recorded in the residual abnormal segment numbers, first read the sample number and start and end time points of each set of data, such as (P24, 1113), and then call the unique identification code of the sample in the structure database for association. For example, the structure number corresponding to P24 in the sampling table is "ZL-P24-BT1". By looking up its original monitoring log, all monitoring records under this number are confirmed, and the deformation and moisture change logs within the abnormal time period are extracted. The number and time period are bound. The binding result format is "ZL-P24-BT1: 11h13h". All records are archived once in this way, and the integration results are output in a standard text structure or table format. For example, a list of abnormal deformation sections in CSV format is used. Each record in the list includes fields such as the sample's unique number, the start and end time points of the abnormality, and whether it spans multiple cycles (such as overnight drying). An example is as follows: "ZL-P24-BT1, 11, 13, no". If the abnormal section lasts for more than 8 hours, a warning mark must be added in the remarks. In actual scenarios, the list can be imported into the database for automatic matching of drying line equipment parameter corrections or for subsequent traceability and review operations. The integration of abnormal sections must ensure data integrity, and there must be no number omissions or time period overlaps. A number matching verification mechanism is executed in the merge operation. Once repeated sample numbers or overlapping time periods are found, the process must be forcibly interrupted for reminder verification.

[0033] See also Figure 6 , the steps to obtain S5 are: S501: Calling samples from the list of abnormal deformation sections, extracting the water gradient direction change values ​​and structural angle change rate values ​​for multiple stages, synchronously integrating the corresponding data by time point, and establishing a stage gradient and angle sequence set; When calling the samples in the deformation abnormal section list, each abnormal sample number and its corresponding abnormal time period are obtained in turn. For each sample number, the moisture gradient direction change value and structural angle change rate value recorded in each stage during the time period are extracted from the corresponding drying process data. For example, the sample number is ZL-G17-D2, and its abnormal time period is from the 10th to the 14th hour. In the corresponding drying data, the surface moisture content at the 10th hour is 20.8%, and the core layer is 15.3%. The calculated gradient value is (20.8-15.3) / 5=1.1% / cm. At the 11th hour, they are 19.5% and 14.9% respectively, and the gradient change is 0.92% / cm. Then continue to extract until the 14th hour, and the sequence obtained is {1.1, 0.92, 0.85, 0.79, 0.70}. The structural angle change rate value is obtained by differential angle measurement data. Assuming that the angle at the 10th hour is The first hour was 3.2°, the 11th hour was 3.6°, and the rate was (3.6-3.2) / 1=0.4° / h, resulting in a continuous rate sequence {0.4, 0.38, 0.35, 0.31, 0.28}. All data were based on a unified sampling time interval of 1 hour. A data structure with one-to-one correspondence between moisture gradient changes and structural angle rates was established for sample number ZL-G17-D2. Each set of data records contained three parameters: time point, moisture gradient change value, and angle rate value. To ensure the accuracy of data synchronization, the system compared the timestamps. If data was missing or the recording time was inconsistent, the sample was discarded to ensure the reliability of subsequent analysis. The unit of moisture gradient change value in each stage was set as percentage per centimeter, and the unit of structural angle rate was angle per hour. All units were uniformly converted, and the data was stored in a table structure and in CSV format for subsequent calls and trend analysis.

[0034] S502: Calculate the offset trend value between the moisture gradient direction change and the angle change rate based on the data of multiple time points in the stage gradient and angle sequence, determine whether it falls within the set trend deviation range, select the sample numbers that meet the conditions, and obtain the trend offset mark sequence; According to the data of multiple time points in the stage gradient and angle sequence, each set of data is compared and processed to extract the synchronous offset trend value between the moisture gradient direction change and the angle change rate. The specific implementation method is to take out the moisture gradient change value and the structural angle rate value at each time point. For example, the gradient of the ZL-G17-D2 sample is 1.1% / cm and the angle rate is 0.40° / h at the 10th hour, and 0.92% / cm and 0.38° / h at the 11th hour respectively. The offset trend of this stage is the difference between the gradient and the angle rate before and after, that is, |(1.1-0.92)-(0.40-0.38)|=0.16-0.02=0.14. The offset trend value is recorded as 0.14. Similarly, the trend sequence {0.14, 0.10, 0.08, 0.07} is obtained. Then the judgment operation is performed to screen whether there is a significant trend offset based on the set trend offset threshold range. The threshold setting interval is ±0.12. This range is Based on the analysis of 120 groups of normal and abnormal drying samples in previous experiments, multiple comparisons found that when the absolute value of the trend offset value exceeded 0.12, the probability of the corresponding structure having transverse cracks or tensile deformation reached 72%. For samples with an absolute value below 0.12, this probability dropped to below 15%. Therefore, ±0.12 was set as the trend judgment interval. During screening, each offset trend value was judged to determine whether it exceeded this interval. For example, a trend value of 0.14 exceeding 0.12 was judged as "offset", and a value of 0.10 within the interval was judged as "normal". If the trend value offset at any two or more consecutive time points was judged as "offset", the entire sample was marked as a trend offset sample. All samples that met the conditions were numbered into a trend offset label sequence, such as {ZL-G17-D2, ZL-K05-F1, ZL-M88-C3}. Each number in this sequence represents a sample that has been determined to have an abnormal structural trend response, which is used for further processing and decision-making.

[0035] S503: Calling the offset sample number in the trend offset mark sequence, grouping the corresponding samples and adjustment schemes, building a correspondence between the sample number and the processing structure, and obtaining the medicinal material drying processing scheme; When calling the offset sample number in the trend offset mark sequence, first read the sample numbers in the sequence one by one, and search for the historical information and physical classification information of the corresponding sample by number. For example, ZL-G17-D2 belongs to the "root" medicinal material. According to its category and the response recorded in the corresponding historical drying data, search for the adjustment strategy that has been applied or recommended in the drying adjustment strategy database. Combined with the trend offset response, a recommended adjustment plan is constructed. For example, in the trend offset, the drying gradient of this sample decreases faster than the structural response, which belongs to the "too fast moisture migration" type problem. The corresponding processing strategy is "cooling and deceleration to extend the stage II time for 2 hours". The sample and the adjustment plan are combined into "ZL-G17-D2: extension II + 2h". All offset samples perform the same process in sequence to establish a one-to-one correspondence between the sample number and the processing structure. All control structures are organized into a list of medicinal material drying treatment plan structures. The list has a unified format and the fields are "sample number", "herb category", and "recommended strategy". Each recommended strategy must clearly define the adjustment target, parameter adjustment direction and adjustment range. Ambiguous terms such as "appropriate increase" or "appropriate decrease" must not appear. Instead, the value must be specified, such as "extend the time by 2 hours" or "lower the temperature by 5°C". In the process of formulating the treatment plan, three adjustment ranges have been established for each type of medicinal material. Among them, mild deviation corresponds to an adjustment range of ≤1 hour, moderate is 1~2 hours, and severe is ≥2 hours. ZL-G17-D2 is a "moderate deviation" with a maximum trend value deviation of 0.14, so its adjustment recommendation is set to "extend II + 2h". All results are summarized in the "Drying Treatment Plan Table" for system execution or manual review.

[0036] The medicinal material drying system based on big data includes: The moisture grouping module obtains the surface and center moisture values ​​of the sample, constructs a time series, calculates the rate of change, extracts the rate difference sequence between samples, determines whether it exceeds the set standard, classifies the sample pairs, and obtains a set of difference group numbers; The trend configuration module divides the samples based on the difference group number set, extracts the rate sequence, calculates the average trend, determines the segment to which it belongs, matches the set number, combines the ownership information, and obtains the starting configuration matching table; The origin offset module extracts the origin from the starting configuration matching table, collects humidity and temperature parameters, obtains the drying inflection point and end time, calculates and sorts the time difference, performs index difference comparison, and obtains the origin offset sequence set; The deformation identification module extracts offset samples based on the origin offset sequence set, calls the moisture gradient map and concentration ratio sequence, pairs the structural angle values, constructs a residual sequence, determines whether the error limit is continuously exceeded, records abnormal samples and time periods, and obtains a list of abnormal deformation sections; The adjustment scheme module calls the list of abnormal deformation sections, extracts the moisture gradient direction and angle change rate, calculates the offset trend, determines whether it enters the deviation range, groups the sample numbers, and obtains the medicinal material drying treatment plan.

[0037] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A medicinal material drying method based on big data, characterized in that: The following steps are involved: S1: Obtain the moisture values ​​of the surface and center of the Chinese medicinal material samples, construct a time series based on the sampling interval, calculate the moisture change rate at adjacent time points, extract the rate difference sequence between samples, perform a comparison to see if they exceed the judgment threshold, classify the samples pairwise, and obtain a difference group number set; S2: Divide the samples according to the difference group number set, extract the rate sequence to calculate the average trend, determine the segment to which the trend belongs, match the set number interval, combine the corresponding number and sample ownership information, and obtain the initial configuration matching table; S3: Extracting the sample origins from the initial configuration matching table, collecting humidity, temperature frequency, and day / night difference, obtaining the inflection point and end time in the drying curve, calculating the humidity and inflection point time difference, and the end time offset, constructing two sorted sequences, and performing index difference comparison to obtain an origin offset sequence set; S4: Extract the offset samples from the origin offset sequence set, call the moisture gradient map and concentration ratio sequence of the drying process, pair the structural deformation angle sequence, construct the residual sequence at the same time point, determine whether it exceeds the error limit continuously, record the abnormal number and time period, and obtain a list of deformation abnormal sections.

2. The medicinal material drying method based on big data according to claim 1, characterized in that: The difference group number set includes the moisture change rate difference, sample classification number, and rate difference threshold identifier; the starting configuration matching table includes the group number correspondence, the trend segment number to which the sample belongs, and the configuration matching number; the origin offset sequence set includes the humidity time offset, the end time offset, and the origin sorting index difference; the deformation abnormality segment list includes the abnormal sample number, the abnormal time period number, and the residual over-limit mark.

3. The medicinal material drying method based on big data according to claim 2, characterized in that: The steps for obtaining S1 are: S101: Based on the time points at which the surface moisture value and the center moisture value of the Chinese medicinal material sample are obtained, the change in moisture value at adjacent time points is calculated, and the difference rate between the surface moisture value and the center moisture value at adjacent time points is determined. By constructing a time series, the trend of moisture difference change at consecutive time points is obtained, and a moisture difference change rate sequence is generated; S102: Calling the rate values ​​at the same time between any two samples in the moisture difference change rate sequence, constructing a time series difference degree based on the time point difference change, calculating and obtaining a rate difference index between samples, and establishing a sample rate difference index set; S103: Based on the difference index value between any two samples in the sample rate difference index set, call the set moisture change difference threshold, judge whether the difference value exceeds the threshold in turn, classify according to whether it exceeds, divide the samples that meet the same classification conditions into a unified number group, and obtain a difference group number set.

4. The medicinal material drying method based on big data according to claim 3, characterized in that: The steps for obtaining S2 are: S201: Divide the Chinese medicinal material samples based on the differential group number set, call the moisture change rate sequence corresponding to the sample under the number, perform a sum operation on the sample rate values ​​according to the time dimension and divide it by the number of samples, obtain the average rate trend value sequence of the time points under the differential grouping, and generate a group average trend sequence; S202: Calling the time series mean and series slope under multiple numbers in the grouped average trend sequence, calculating and obtaining the trend index value under multiple numbers, calling the set segment standard according to the position of the trend index value within the segment range, determining the segment number to which it belongs, and generating a trend segment number sequence; S203: According to the trend segment number sequence, each number is combined with the corresponding sample affiliation number and group number to establish a corresponding structure of sample affiliation, trend segment and group number to obtain an initial configuration matching table.

5. The medicinal material drying method based on big data according to claim 4, characterized in that: The steps for obtaining S3 are: S301: Obtain the sample origin in the initial configuration matching table, collect humidity, temperature frequency and day and night difference, extract the inflection point time and end time in the drying curve, calculate the humidity and inflection point time difference and make corrections to obtain the drying process climate offset value sequence; S302: calling the inflection point time difference and the end time offset value in the drying process climate offset value sequence, constructing two sorting sequences and extracting index numbers, calculating the difference between the indexes, and obtaining a sorting index offset sequence; S303: Rearranging the order of the samples according to the index offset value of each origin in the sorted index offset sequence in combination with the sample origin number information, establishing a corresponding structure sequence of samples and origins, and obtaining an origin offset sequence set.

6. The medicinal material drying method based on big data according to claim 5, characterized in that: The acquisition steps of S4 are: S401: extracting the offset samples from the origin offset sequence set, calling the corresponding drying process moisture gradient map and concentration ratio sequence, pairing the structural deformation angle sequence, performing an operation on the difference between the moisture gradient value and the structural angle value at the same time point, constructing a difference set at the time point, and obtaining a residual sequence at the same time point; S402: Calling the time period difference in the residual sequence at the same time point, comparing it with the structural error limit, screening the time periods that continuously exceed the limit, recording the corresponding sample numbers and continuous time period information, and obtaining the residual abnormal segment number set; S403: According to the sample numbers and corresponding time periods recorded in the residual abnormal segment number set, the offset sample identifiers and abnormal time periods are integrated, a corresponding structure between numbers and time periods is established, and a list of abnormal deformation sections is obtained.

7. The medicinal material drying method based on big data according to claim 6, characterized in that: The method further comprises: S5: Calling samples from the abnormal deformation section list, extracting the phase moisture gradient direction change and angle change rate, calculating the offset trend value of the two, determining whether they enter the trend deviation range, grouping them into an adjustment result set, and obtaining a medicinal material drying treatment plan; The medicinal material drying treatment scheme includes a moisture gradient direction offset value, an angle change rate offset value, and a trend deviation identification result; The steps to obtain S5 are: S501: calling samples in the abnormal deformation section list, extracting the moisture gradient direction change values ​​and structural angle change rate values ​​of multiple stages, synchronously integrating the corresponding data according to time points, and establishing a stage gradient and angle sequence set; S502: Calculate the offset trend value between the moisture gradient direction change and the angle change rate based on the data of multiple time points in the stage gradient and angle sequence, determine whether it falls within the set trend deviation range, select the sample numbers that meet the conditions, and obtain the trend offset mark sequence; S503: calling the offset sample number in the trend offset mark sequence, grouping the corresponding samples and adjustment schemes, building a correspondence between the sample number and the processing structure, and obtaining the medicinal material drying processing scheme.

8. The medicinal material drying system based on big data is characterized by: The system is used to execute the medicinal material drying method based on big data according to any one of claims 1 to 7, comprising: The moisture grouping module obtains the surface and center moisture values ​​of the sample, constructs a time series, calculates the rate of change, extracts the rate difference sequence between samples, determines whether it exceeds the set standard, classifies the sample pairs, and obtains a set of difference group numbers; The trend configuration module divides the samples based on the difference group number set, extracts the rate sequence, calculates the average trend, determines the segment to which it belongs, matches the set number, combines the attribution information, and obtains the starting configuration matching table; The origin offset module extracts the origin in the starting configuration matching table, collects humidity and temperature parameters, obtains the drying inflection point and end time, calculates and sorts the time difference, performs index difference comparison, and obtains an origin offset sequence set; The deformation identification module extracts offset samples based on the origin offset sequence set, calls the moisture gradient map and concentration ratio sequence, pairs the structural angle values, constructs a residual sequence, determines whether the error limit is continuously exceeded, records abnormal samples and time periods, and obtains a list of abnormal deformation sections; The adjustment scheme module calls the abnormal deformation section list, extracts the moisture gradient direction and angle change rate, calculates the offset trend, determines whether it enters the deviation range, groups the sample numbers, and obtains the medicinal material drying treatment plan.

Images