Traditional Chinese medicine drying quality control method and system based on artificial intelligence

Abstract

The invention relates to the technical field of traditional Chinese medicine drying, in particular to a traditional Chinese medicine drying quality control method and system based on artificial intelligence, and the method comprises the steps: measuring the temperature of drying airflow introduced into a drying tower, the concentration of to-be-dried feed liquid, and the atomization pressure and drying time for atomizing the to-be-dried feed liquid; a drying quality prediction model is established, the temperature of the drying airflow, the concentration of the to-be-dried feed liquid, the atomization pressure and the drying time are input into the drying quality prediction model, and a drying quality prediction value is obtained; determining a drying quality target value, and analyzing the drying quality predicted value and the drying quality target value to obtain a drying quality difference result; and the drying quality difference result is input into the quality control model, drying adjustment parameters are obtained, and all drying parameters are adjusted according to the drying adjustment parameters. According to the method and system, the traditional Chinese medicine drying quality can be stabilized, and the reject ratio of all-batch dried products is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of traditional Chinese medicine drying, and in particular to an artificial intelligence-based traditional Chinese medicine drying quality control method and system. Background Art

[0002] Traditional Chinese medicine spray drying is an effective technology commonly used to prepare traditional Chinese medicine solid dispersions. It quickly evaporates the water in the liquid to be dried in a high-temperature airflow to obtain traditional Chinese medicine powder or granular material. This method can effectively retain the active ingredients of traditional Chinese medicine and is suitable for preparing traditional Chinese medicine granules, powders, oral liquids and other preparations.

[0003] The traditional spray drying process of traditional Chinese medicine usually relies on manual control, relying on experience judgment and timing operation to adjust the drying parameters. Various influencing factors in the drying process (such as air flow temperature, humidity changes, spray pressure, liquid concentration, etc.) are highly complex and nonlinear. It is difficult to accurately grasp the changes in key quality indicators by relying on experience judgment, resulting in unstable drying quality. At the same time, the discovery of quality abnormalities has a lag, and the drying parameters cannot be adjusted in a timely and effective manner, resulting in a high rate of defective quality in some batches. Summary of the invention

[0004] To this end, the technical problem to be solved by the present invention is to overcome the problems of unstable spray drying quality and high defective rate of some batches in the prior art, and to provide a Chinese medicine drying quality control method and system based on artificial intelligence to stabilize the Chinese medicine drying quality and effectively improve the defective rate of the entire batch of dried products.

[0005] In the first aspect, in order to solve the above technical problems, the present invention provides a method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence, comprising:

[0006] Passing a dry air flow into the drying tower, and measuring the temperature of the dry air flow;

[0007] Determine the concentration of the liquid to be dried; atomize the liquid to be dried, allow the atomized liquid to enter the drying tower, and measure the atomization pressure of the liquid to be dried;

[0008] Drying the atomized liquid to be dried by the drying airflow in the drying tower, and measuring the drying time;

[0009] Establishing a drying quality prediction model, inputting the temperature of the drying airflow, the concentration of the liquid to be dried, the atomization pressure and the drying time into the drying quality prediction model to obtain a drying quality prediction value; wherein the drying quality indicators include moisture content, active ingredient content, particle size and bulk density;

[0010] Determine a drying mass target value, and analyze the drying mass prediction value and the drying mass target value to obtain a drying mass difference result;

[0011] The drying quality difference result is input into a quality control model to obtain a drying adjustment parameter, and the temperature of the drying airflow, the atomization pressure, the concentration of the liquid to be dried and the drying time are adjusted according to the drying adjustment parameter.

[0012] In one embodiment of the present invention, a drying quality prediction model is established, comprising:

[0013] Obtain historical data information of the drying process of the same type of liquid as the liquid to be dried, and extract sample data information related to the drying quality;

[0014] Construct a sample data window, and configure weights for the sample data information in the sample data window; wherein the sample data window is divided into a plurality of time points according to a time sequence, and weights for the sample data information at each time point are configured according to formula (1):

[0015]

[0016] i represents the time point number; f(i) represents the weight of the i-th time point; ti represents the time difference between the current time point and the data collection time point; λ represents the decay rate parameter, which affects the speed at which the weight decays over time;

[0017] The sample data information is deeply learned according to the configuration weights to obtain the drying quality prediction model.

[0018] In one embodiment of the present invention, after configuring the weights according to formula (1), the method further includes normalizing the configured weights according to the following method:

[0019]

[0020] Among them, f(i)' represents the normalized weight of the i-th time point; n represents the number of time points; j represents the sequence number of the divided time points; f(j) represents the weight of the j-th time point;

[0021] The sample data information is deeply learned according to the weights after the normalization operation to obtain the drying quality prediction model.

[0022] In one embodiment of the present invention, the drying quality prediction model includes an input layer, an LSTM layer, a fully connected layer and an output layer;

[0023] The sample data information of the sample data window is converted into a feature matrix capable of performing convolution operation at the input layer;

[0024] The feature matrix is ​​subjected to feature extraction at the LSTM layer, and the LSTM layer can identify abnormal quality features of sample data;

[0025] The feature extraction result is mapped and transformed in the fully connected layer;

[0026] Outputting the drying quality prediction value through the output layer;

[0027] The LSTM layer sets an activation function, and performs nonlinear transformation on the feature extraction result through the activation function.

[0028] In one embodiment of the present invention, the control method further includes:

[0029] Collect the actual measured value of drying mass;

[0030] Calculating the difference between the actual dry mass value and the predicted dry mass value to obtain a predicted difference result;

[0031] The model parameters of the drying mass prediction model are adjusted according to the prediction difference result.

[0032] In one embodiment of the present invention, the model parameters of the drying mass prediction model are adjusted according to the prediction difference result, including:

[0033] Performing difference distribution analysis and error deviation trend analysis on the predicted difference results to obtain predicted difference analysis results;

[0034] Determining the number of neurons in the LSTM layer according to the activation degree of the LSTM layer;

[0035] The number of neurons in the LSTM layer is adjusted according to the prediction difference analysis results.

[0036] In one embodiment of the present invention, determining the number of neurons in the LSTM layer according to the activation degree of the LSTM layer includes:

[0037] For the LSTM layer, collect the activation function output value of each neuron to obtain an activation value set;

[0038] Calculating at least one of a maximum value, a minimum value, an average value and a standard value for the activation value set;

[0039] The number of neurons in the LSTM layer is determined according to the calculation results.

[0040] In one embodiment of the present invention, the control method further includes:

[0041] Collect the actual measured value of drying mass;

[0042] Analyzing the actual drying mass value and the target drying mass value to obtain an operation difference result;

[0043] The drying adjustment parameter is adjusted according to the operation difference result.

[0044] In one embodiment of the present invention, the drying adjustment parameter is adjusted according to the operation difference result, including:

[0045] Calculating the difference between the actual dry mass value and the target dry mass value;

[0046] The running variance rate is calculated as follows:

[0047]

[0048] Among them, α represents the operation difference rate; M2 represents the actual value of drying quality; M1 represents the target value of drying quality;

[0049] If the operation difference rate is less than 5%, the current drying adjustment parameters are maintained;

[0050] If the operation difference rate is greater than or equal to 5%, the operation difference rate and the actual drying quality measurement value are input into the quality control model to adjust the drying adjustment parameter.

[0051] In a second aspect, in order to solve the above-mentioned technical problems, the present invention provides a Chinese medicine drying quality control system based on artificial intelligence, which is used to execute the above-mentioned Chinese medicine drying quality control method based on artificial intelligence.

[0052] The above technical solution of the present invention has the following beneficial effects compared with the prior art:

[0053] The artificial intelligence-based Chinese medicine drying quality control method and system described in the present invention combines multiple key parameters such as the temperature of the drying airflow, the concentration of the liquid to be dried, the atomization pressure and the drying time, and uses artificial intelligence technology to construct a drying quality prediction model, which can predict the quality changes in the drying process in multiple dimensions; by analyzing the drying quality difference results, the key parameters are adjusted to ensure that the quality control of the drying process is more accurate and efficient; at the same time, a deep learning model is used to achieve dynamic optimization of the Chinese medicine drying process and early warning of quality anomalies, solve the problems of unstable spray drying quality and high defective rate of some batches, stabilize the drying quality of Chinese medicine, and dry the entire batch product The defective rate has been effectively improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to make the content of the present invention more clearly understood, the present invention is further described in detail below according to specific embodiments of the present invention in conjunction with the accompanying drawings, wherein:

[0055] Figure 1 It is a flow chart of a method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence in a preferred embodiment of the present invention;

[0056] Figure 2 A flow chart of establishing a drying quality prediction model in a preferred embodiment of the present invention;

[0057] Figure 3 It is a framework diagram of a drying quality prediction model in a preferred embodiment of the present invention; DETAILED DESCRIPTION

[0058] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments so that those skilled in the art can better understand the present invention and implement it, but the embodiments are not intended to limit the present invention.

[0059] Embodiment 1

[0060] Reference Figure 1 As shown, the embodiment of the present invention discloses a method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence, comprising:

[0061] S100, introducing a dry air flow into a drying tower, and measuring the temperature of the dry air flow;

[0062] S200, measuring the concentration of the liquid to be dried; atomizing the liquid to be dried, allowing the atomized liquid to be dried to enter the drying tower, and measuring the atomization pressure of the liquid to be dried;

[0063] S300, drying the atomized liquid to be dried by the drying airflow in the drying tower, and measuring the drying time;

[0064] In specific application scenarios, the process of spray drying of traditional Chinese medicine includes:

[0065] Spray drying is to spray liquid material (referred to as "liquid") into fine droplets (i.e. "atomization") through a nozzle, and then contact it with a high-temperature airflow (usually hot air), hereinafter referred to as "drying airflow". Under the action of the airflow, the water evaporates rapidly and finally forms a powder. The whole process usually includes the following steps:

[0066] Generate dry air flow: introduce dry air flow (usually air at 180°C-220°C) into the drying tower through an air heater and an air conveyor, and measure the temperature of the dry air flow in real time during the introduction of the dry air flow.

[0067] Pretreatment: The liquid to be dried is filtered, concentrated, and the concentration is adjusted to make it suitable for spray drying; after the pretreatment of the liquid to be dried is completed, the concentration of the liquid is measured.

[0068] Spraying process: The pretreated liquid to be dried is atomized by a spray device (such as a spray tower or a spray disk) into fine droplets that enter the drying tower. The atomization pressure is measured during this process.

[0069] Drying process: When tiny droplets come into contact with the dry air flow, the water in the droplets is quickly evaporated, and the remaining solid components form powder; as the water in the droplets evaporates, the powder will quickly condense into tiny particles and eventually be collected at the bottom of the drying tower. The drying time is measured during this process.

[0070] S400, establishing a drying quality prediction model, inputting the temperature of the drying airflow, the concentration of the liquid to be dried, the atomization pressure and the drying time into the drying quality prediction model, and obtaining a drying quality prediction value; wherein the drying quality indicators include moisture content, active ingredient content, particle size and bulk density;

[0071] In specific application scenarios, the goal of the drying quality prediction model is to predict the final drying quality (drying quality indicators include moisture content, active ingredient content, particle size and bulk density) based on the key parameters of the drying process (including the temperature of the drying airflow, the concentration of the liquid to be dried, the atomization pressure and the drying time). In order to establish an effective drying quality prediction model, it is necessary to first collect historical data related to the drying process and extract sample data related to the drying quality from the historical data, including process parameter data, such as the drying airflow temperature in each drying cycle, the concentration of the liquid to be dried before each drying, the atomization pressure in each atomization process and the drying time of each drying process; corresponding to the quality evaluation data of each drying cycle. These data can be collected in real time through online sensors and laboratory analysis equipment. In order to improve the accuracy of the model, the collected data should include multiple drying cycles to cover different drying process conditions. A deep learning model (such as long short-term memory network LSTM, deep neural network, convolutional neural network, etc.) is used to perform deep learning on historical data to obtain a drying quality prediction model. The drying quality prediction model that has been trained and verified to be effective is put into actual production use. The drying air flow temperature, concentration of the liquid to be dried, atomization pressure and drying time in the current drying process are input into the drying quality prediction model to obtain the predicted values ​​of various drying quality indicators.

[0072] S500, determining a drying mass target value, and analyzing the drying mass prediction value and the drying mass target value to obtain a drying mass difference result;

[0073] In specific application scenarios, the setting of drying quality target values ​​not only needs to consider the quality standards of traditional Chinese medicine, but also needs to be combined with factors such as production process, equipment capacity, and medicinal material characteristics; the target value generally refers to the ideal or standard value of key quality indicators in the drying process (such as moisture content, active ingredient content, particle size, and bulk density). It will serve as the basis for drying quality control and be used to compare with the actual predicted value, thereby providing guidance for the adjustment of the production process. Among them, the drying quality target value can be determined according to a variety of methods, such as determining the target value through historical data; the target value can also be determined in combination with pharmacopoeia and standardization requirements, and the target value can be determined through experiments.

[0074] The drying mass target value is a fixed value, and the drying mass difference result can be obtained by analyzing the difference between the drying mass predicted value and the drying mass target value.

[0075] S600, inputting the drying quality difference result into a quality control model to obtain drying adjustment parameters, and adjusting the temperature of the drying airflow, the atomization pressure, the concentration of the liquid to be dried and the drying time according to the drying adjustment parameters.

[0076] In specific application scenarios, the drying quality difference results are input into the quality control model. The quality control model should be based on the characteristics and requirements of the drying process of the liquid to be dried, and establish an adaptive control model for each process of the drying process. For example, it can be a model based on neural network models, genetic algorithms and other methods. According to the model calculation and analysis results, the optimal drying parameters are obtained. The drying parameters include drying airflow temperature, liquid concentration, atomization pressure, drying time, etc. The best drying quality indicators are obtained after running according to these drying parameters. For example, if the moisture content is too high, it may be due to the spray drying temperature being too low or the drying time being too short. At this time, the moisture content can be adjusted by adjusting the temperature, wind speed, drying time and other parameters of the spray drying tower; if the content of the active ingredient is low, the extraction concentration of the raw material can be adjusted or the extraction process can be optimized; if the particle size distribution is uneven, it may be due to unstable nozzle pressure or too large droplet size. Adjusting the spray pressure and droplet size will help the uniformity of the particle size; if the bulk density is not suitable, it can be optimized by changing the compaction method of the powder or increasing the vibration density of the material.

[0077] The artificial intelligence-based Chinese medicine drying quality control method described in the present invention combines multiple key parameters such as the temperature of the drying airflow, the concentration of the liquid to be dried, the atomization pressure and the drying time, and uses artificial intelligence technology to construct a drying quality prediction model, which can predict the quality changes in the drying process in multiple dimensions; by analyzing the drying quality difference results, the various key parameters are adjusted to ensure that the quality control of the drying process is more accurate and efficient; at the same time, a deep learning model is used to achieve dynamic optimization of the Chinese medicine drying process and early warning of quality anomalies, solve the problems of unstable spray drying quality and high defective rate of some batches, stabilize the Chinese medicine drying quality, and effectively improve the defective rate of the entire batch.

[0078] Further, refer to Figure 2 As shown in the figure, a drying quality prediction model is established, including:

[0079] Acquire historical data information of the drying process of the liquid to be dried, and extract sample data information related to drying quality;

[0080] In specific application scenarios, historical data information of each link in the drying process of the same type of liquid to be dried is obtained, such as temperature, humidity, spray pressure, drying air flow rate, drying time, spray volume, etc.; quality indicator data, such as particle size, moisture content, effective ingredient content, bulk density; and external environment data, such as control humidity, temperature, feeding conditions, etc.; data related to drying quality evaluation indicators (for example) are extracted from these historical data information, and these data are used as sample data for drying quality prediction model training.

[0081] Construct a sample data window, and configure weights for the sample data information in the sample data window; wherein the sample data window is divided into a plurality of time points according to a time sequence, and weights for the sample data information at each time point are configured according to formula (1):

[0082]

[0083] i represents the time point number; f(i) represents the weight of the i-th time point; ti represents the time difference between the current time point and the data collection time point; λ represents the decay rate parameter, which affects the speed at which the weight decays over time;

[0084] The sample data information is deeply learned according to the configuration weights to obtain the drying quality prediction model.

[0085] In specific application scenarios, a sample data window is constructed according to the time sequence. The sample data window is designed to collect sample data within a period of time (for example, every hour or every 5 minutes). The data in each time window contains the drying process parameters and quality indicators within a period of history. Assuming that the data window is T hours, all sample data in each T-hour window (such as drying air flow temperature, humidity, spray pressure, moisture content, etc.) constitute a data set. The extracted sample data information is arranged in chronological order for subsequent learning and analysis. According to specific needs and actual conditions, the capacity and time length of the sample data window are set. The capacity can represent the number of items of sample data information stored in the window, and the time length represents the time span of each data entry. According to the set time step, the sample data window is updated, that is, the data information in the sample data window is updated regularly to keep the data in the window in a certain timeliness with the actual situation.

[0086] The data in the sample data window are weighted in chronological order. In the drying process, the most recent operating parameters usually have a greater impact on the current quality, while the more distant historical data have a smaller impact; based on this, the weights are configured according to a specific exponential decay method to emphasize that the closer the data is to the current time, the greater the weight is configured; that is, the data at the most recent moment is assigned a higher weight, while the data far from the current time is assigned a lower weight. In the scheme of the embodiment of the present invention, by introducing time series weights, the drying quality prediction model can more accurately capture the impact of recent data, thereby improving the prediction accuracy of the drying quality change trend.

[0087] Furthermore, after configuring the weights according to formula (1), the method further includes normalizing the configured weights in the following manner, performing deep learning on the sample data information according to the weights after the normalization operation, and obtaining the drying quality prediction model.

[0088]

[0089] Among them, f(i)' represents the normalized weight of the i-th time point; n represents the number of time points; j represents the sequence number of the divided time points; f(j) represents the weight of the j-th time point; As the denominator of normalization, it ensures that the configuration weights at each time point still maintain their relative proportions after normalization, and that the sum of all weights equals 1.

[0090] In the solution of the embodiment of the present invention, by configuring the weight normalization operation, it can be ensured that the sum of all weights is 1, making the weight distribution more reasonable and avoiding the influence of some weights being too large or too small on the model training effect.

[0091] Specifically, in order to achieve high-precision prediction and control of the drying process of traditional Chinese medicine, the embodiment of the present invention proposes a drying quality prediction model based on deep learning. The model is specially designed with a multi-layer structure, including an input layer, an LSTM layer, a fully connected layer, and an output layer, so as to make full use of the characteristics of time series data and perform effective feature extraction and prediction. Figure 3 As shown, the drying quality prediction model includes an input layer, an LSTM layer, a fully connected layer and an output layer; the input layer receives sample data information as input, which is usually represented as a multidimensional array; the sample data information of the sample data window is converted into a feature matrix capable of performing convolution operations in the input layer; the feature matrix performs feature extraction in the LSTM layer, and the LSTM layer can identify abnormal quality features of the sample data; the feature extraction result is mapped and transformed in the fully connected layer; the drying quality prediction value is output through the output layer; wherein the LSTM layer sets an activation function, and the feature extraction result is nonlinearly transformed through the activation function.

[0092] In specific application scenarios, the data at each time point (such as drying air flow temperature, concentration of the liquid to be dried, atomization pressure, drying time, moisture content, active ingredient content, bulk density and particle size, etc.) are converted into feature matrices in the input layer. These feature matrices not only contain the original data, but can also be further optimized through additional preprocessing steps (such as normalization). Feature extraction is performed in the LSTM layer to identify abnormal quality features in the sample data; activation functions (such as tanh, ReLU, etc.) are set to enhance the expressive power of the model through nonlinear transformation. For example, using tanh as an activation function can help the LSTM layer better handle the balance between negative and positive values, while ReLU helps to accelerate the training process. The fully connected layer maps and transforms the feature extraction results output by the LSTM layer to provide a compact and effective feature vector for the final prediction. The key prediction indicators of drying quality, such as moisture content, active ingredient content, particle size and bulk density, are directly given through the output layer. The powerful time series processing capabilities of the LSTM layer are combined with the feature mapping of the fully connected layer to improve the accuracy and reliability of drying quality prediction.

[0093] Furthermore, the control method also includes: collecting the actual measured value of the drying mass; calculating the difference between the actual measured value of the drying mass and the predicted value of the drying mass to obtain a predicted difference result; and adjusting the model parameters of the drying mass prediction model according to the predicted difference result.

[0094] Adjusting the model parameters of the drying quality prediction model specifically includes: performing difference distribution analysis and error deviation trend analysis on the prediction difference results to obtain prediction difference analysis results; determining the number of neurons in the LSTM layer according to the activation degree of the LSTM layer; and adjusting the number of neurons in the LSTM layer according to the prediction difference analysis results.

[0095] In specific application scenarios, difference distribution analysis is used to evaluate the deviation and error characteristics between the predicted value of dry mass and the measured value of dry mass. For each time step, the difference between the two is used to obtain the error value. The error values ​​of all time steps are statistically analyzed to obtain the difference distribution. According to the difference distribution, it is analyzed whether the error is concentrated in certain specific ranges or shows large fluctuations. Error deviation trend analysis is used to identify the law of change of prediction error over time, whether it shows a trend of gradual increase or decrease, or whether there are periodic fluctuations. This trend may indicate that the model fails to fully capture the law of certain time series data, resulting in large prediction errors in specific time periods. The activation degree of the LSTM layer can be measured by calculating the output value of the activation function (such as the output of the tanh or sigmoid function). If the output of most neurons is close to 0 or 1, it means that the activation degree is too low or too high, which may lead to gradient vanishing or gradient explosion. The number of neurons in the LSTM layer determines the expression ability and fitting ability of the network. According to the results of the prediction difference analysis, the number of neurons in the LSTM layer is dynamically adjusted: if the error is large and the activation degree of the LSTM layer is low, increase the number of neurons in the LSTM layer to improve the complexity of the model so that it can learn more features. If the prediction error shows an unstable trend (such as large fluctuations in the error), increasing the number of neurons can help the network better learn the patterns and regularities of the input data. If the error is small and stable, and the activation level is high, keep the current number of neurons to avoid overfitting the model.

[0096] Specifically, according to the activation degree of the LSTM layer, the number of neurons in the LSTM layer is determined, including collecting the activation function output value of each neuron in the LSTM layer to obtain an activation value set; calculating at least one of the maximum value, minimum value, average value and standard value for the activation value set; and determining the number of neurons in the LSTM layer according to the calculation result. The method of determining the number of neurons according to the activation degree can make the model more flexible and adaptable to the needs of different data and tasks. By dynamically adjusting the number of neurons, the performance and generalization ability of the model can be improved, and problems such as overfitting or underfitting can be avoided.

[0097] In other embodiments of the present invention, the control method further includes collecting actual drying mass measured values; analyzing the actual drying mass measured values ​​and the drying mass target values ​​to obtain operation difference results; and adjusting the drying adjustment parameters according to the operation difference results.

[0098] Drying adjustment parameters should refer to the factors that can be adjusted in the important spray drying process, mainly including drying air flow temperature: affecting the water evaporation rate; atomization pressure: affecting the atomization particle size, and then affecting the drying efficiency and product particle size; concentration of the liquid to be dried: affecting the degree of water evaporation and the preservation of active ingredients during the drying process; drying time: affecting the adequacy of powder drying and moisture control.

[0099] Adjustment of drying airflow temperature: If the error between the actual drying mass value and the drying mass target value is small and within the allowable range, maintain the current temperature unchanged and continue monitoring. If the moisture content is too low, it indicates that the drying temperature is too high, resulting in excessive loss of effective ingredients of traditional Chinese medicine. The airflow temperature needs to be lowered. Lowering the temperature helps slow down the evaporation of water and maintain the stability of the effective ingredients. If the moisture content is too high, it indicates insufficient drying. Increase the airflow temperature to accelerate the evaporation of water and shorten the drying time.

[0100] Adjustment of atomization pressure: If the measured particle size is too large and the target particle size is small, increase the atomization pressure to atomize the liquid to be dried finer and improve the drying efficiency. If the particle size is too small and too fine, reduce the atomization pressure to avoid excessive atomization resulting in powder adhesion or poor fluidity.

[0101] Adjustment of the concentration of Chinese medicine extraction: If the water content is high, but the active ingredients are well preserved, it means that the concentration of the liquid to be dried is too high, resulting in slow evaporation of water. Reducing the liquid concentration will make it easier for the water to evaporate, thereby accelerating the drying process. If the active ingredient content is low, adjust the concentration to better preserve the Chinese medicine ingredients during the drying process and avoid excessive dilution.

[0102] Adjustment of drying time: If the moisture content exceeds the standard, it indicates that the drying time may not be long enough. Increase the drying time to further remove moisture and ensure thorough drying. If the moisture content is too low and a large amount of active ingredients are lost, it indicates that the drying time is too long. Shorten the drying time to avoid over-drying and loss of active ingredients.

[0103] Further, the drying adjustment parameter is adjusted according to the operation difference result, specifically including: calculating the difference between the actual drying mass value and the target drying mass value; and calculating the operation difference rate according to the following method:

[0104]

[0105] Among them, α represents the operation difference rate; M2 represents the actual value of drying quality; M1 represents the target value of drying quality;

[0106] If the operating difference rate is less than 5%, keep the current drying adjustment parameters; if the operating difference rate is greater than or equal to 5%, input the operating difference rate and the actual drying quality value into the quality control model to adjust the drying adjustment parameters. If the operating difference rate is less than 5%, it means that the quality of the drying process is close to the target value, and excessive adjustment may lead to instability in the production process; by setting the drying adjustment parameters to be adjusted only when the difference rate is greater than or equal to 5%, frequent small adjustments are avoided, thereby improving production efficiency and reducing unnecessary interference and adjustment processes. At the same time, avoid fluctuations caused by excessive adjustments and improve quality stability. It can also make the system focus more on obvious quality anomalies and ignore small fluctuations. This strategy helps the system focus on the links that really need adjustment and enhances sensitivity and response speed to abnormal situations.

[0107] The artificial intelligence-based Chinese medicine drying quality control method described in the embodiment of the present invention overcomes the problems of unstable spray drying quality and high defective rate of some batches in the prior art, stabilizes the Chinese medicine drying quality, and effectively improves the defective rate of the entire batch.

[0108] Embodiment 2

[0109] The embodiment of the present invention is based on the same inventive concept as the above embodiment, and discloses an artificial intelligence-based Chinese medicine drying quality control system, which is used to execute the above artificial intelligence-based Chinese medicine drying quality control method. The technical effects that can be achieved are as described in the above embodiment and will not be repeated here.

[0110] In summary, the artificial intelligence-based Chinese medicine drying quality control method and system described in the present invention, combined with multiple key parameters such as the temperature of the drying airflow, the concentration of the liquid to be dried, the atomization pressure and the drying time, uses artificial intelligence technology to construct a drying quality prediction model, which can predict the quality changes in the drying process in multiple dimensions; through the analysis of the drying quality difference results, the adjustment of various key parameters is implemented to ensure that the quality control of the drying process is more accurate and efficient; at the same time, the deep learning model is used to realize the dynamic optimization of the Chinese medicine drying process and the early warning of quality abnormalities, solve the problems of unstable spray drying quality and high defective rate of some batches, stabilize the drying quality of Chinese medicine, and effectively improve the defective rate of the whole batch.

[0111] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0112] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0113] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0115] Obviously, the above embodiments are merely examples for clear explanation and are not intended to limit the implementation methods. For those skilled in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from these are still within the protection scope of the invention.

Claims

1. A method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence, characterized in that: include, Passing a dry air flow into the drying tower, and measuring the temperature of the dry air flow; Determine the concentration of the liquid to be dried; atomize the liquid to be dried, allow the atomized liquid to enter the drying tower, and measure the atomization pressure of the liquid to be dried; Drying the atomized liquid to be dried by the drying airflow in the drying tower, and measuring the drying time; Establishing a drying quality prediction model, inputting the temperature of the drying airflow, the concentration of the liquid to be dried, the atomization pressure and the drying time into the drying quality prediction model to obtain a drying quality prediction value; wherein the drying quality indicators include moisture content, active ingredient content, particle size and bulk density; Determine a drying mass target value, and analyze the drying mass prediction value and the drying mass target value to obtain a drying mass difference result; The drying quality difference result is input into a quality control model to obtain a drying adjustment parameter, and the temperature of the drying airflow, the atomization pressure, the concentration of the liquid to be dried and the drying time are adjusted according to the drying adjustment parameter.

2. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 1, characterized in that: Build a drying quality prediction model, including, Obtain historical data information of the drying process of the same type of liquid as the liquid to be dried, and extract sample data information related to the drying quality; Construct a sample data window, and configure weights for the sample data information in the sample data window; wherein the sample data window is divided into a plurality of time points according to a time sequence, and weights for the sample data information at each time point are configured according to formula (1): i represents the time point number; f(i) represents the weight at the i-th time point; ti represents the time difference between the current time point and the data collection time point; λ represents the decay rate parameter, which affects the speed at which the weight decays over time; The sample data information is deeply learned according to the configuration weights to obtain the drying quality prediction model.

3. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 2, characterized in that: After configuring the weights according to formula (1), the method further includes normalizing the configured weights according to the following method: Among them, f(i)' represents the normalized weight of the i-th time point; n represents the number of time points; j represents the sequence number of the divided time points; f(j) represents the weight of the j-th time point; The sample data information is deeply learned according to the weights after the normalization operation to obtain the drying quality prediction model.

4. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 2, characterized in that: The drying quality prediction model includes an input layer, an LSTM layer, a fully connected layer and an output layer; The sample data information of the sample data window is converted into a feature matrix capable of performing convolution operation at the input layer; The feature matrix is ​​subjected to feature extraction at the LSTM layer, and the LSTM layer can identify abnormal quality features of sample data; The feature extraction result is mapped and transformed in the fully connected layer; Outputting the drying quality prediction value through the output layer; The LSTM layer sets an activation function, and performs nonlinear transformation on the feature extraction result through the activation function.

5. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 4, characterized in that: The control method also includes, Collect the actual measured value of drying mass; Calculating the difference between the actual dry mass value and the predicted dry mass value to obtain a predicted difference result; The model parameters of the drying mass prediction model are adjusted according to the prediction difference result.

6. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 5, characterized in that: The model parameters of the drying quality prediction model are adjusted according to the prediction difference result, including Performing difference distribution analysis and error deviation trend analysis on the predicted difference results to obtain predicted difference analysis results; Determining the number of neurons in the LSTM layer according to the activation degree of the LSTM layer; The number of neurons in the LSTM layer is adjusted according to the prediction difference analysis results.

7. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 6, characterized in that: According to the activation degree of the LSTM layer, the number of neurons in the LSTM layer is determined, include, For the LSTM layer, collect the activation function output value of each neuron to obtain an activation value set; Calculating at least one of a maximum value, a minimum value, an average value and a standard value for the activation value set; The number of neurons in the LSTM layer is determined according to the calculation results.

8. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 1, characterized in that: The control method also includes, Collect the actual measured value of drying mass; Analyzing the actual drying mass value and the target drying mass value to obtain an operation difference result; The drying adjustment parameter is adjusted according to the operation difference result.

9. The method for controlling the drying quality of traditional Chinese medicine based on artificial intelligence according to claim 7, characterized in that: The drying adjustment parameters are adjusted according to the operation difference result, including: Calculating the difference between the actual dry mass value and the target dry mass value; The running variance rate is calculated as follows: Among them, α represents the operation difference rate; M2 represents the actual value of drying quality; M1 represents the target value of drying quality; If the operation difference rate is less than 5%, the current drying adjustment parameters are maintained; If the operation difference rate is greater than or equal to 5%, the operation difference rate and the actual measured value of the drying quality are input into the quality control model to adjust the drying adjustment parameter.

10. A Chinese medicine drying quality control system based on artificial intelligence, characterized in that: Used to execute the artificial intelligence-based traditional Chinese medicine drying quality control method as described in any one of claims 1-9.

Images