Intelligent monitoring system and method for discharging of pharmaceutical production

Abstract

The application discloses a discharging intelligent monitoring system and method for medicine production, and belongs to the field of medicine discharging monitoring. The application obtains discharging characteristic parameters in a powder discharging process, extracts discharging state characteristics based on the obtained discharging characteristic parameters, generates a powder discharging state characteristic vector, obtains reaction kettle production parameters, performs powder discharging risk assessment based on the reaction kettle production parameters and the discharging state characteristics, performs risk influence degree determination based on the powder discharging risk, and performs early warning and regulation, so that the discharging stability is improved, and the production efficiency is improved.

Description

Technical Field

[0001] This application belongs to the field of pharmaceutical discharge monitoring, specifically an intelligent discharge monitoring system and method for pharmaceutical production. Background Technology

[0002] In the pharmaceutical industry, the conveying and discharging of powder materials is a crucial step in the production process. During the transport of powder materials from containers such as reactors, storage tanks, or hoppers to downstream equipment, they are often affected by multiple factors, including material characteristics, the structure of the discharging equipment, vibration, temperature, and process parameters. This can easily lead to abnormalities such as uneven discharging, intermittent flow, bridging, or agglomeration. These abnormalities not only reduce production efficiency but can also cause excessive equipment load, pipeline blockage, and even affect the quality stability of the final product. For example, in the production of solid pharmaceutical dosage forms, powder agglomeration can cause uneven tableting or mixing downstream, affecting the accuracy of drug dosage. Currently, monitoring methods for the powder discharging process mainly rely on manual inspection or simple physical quantity measurements, such as weighing devices, motor current monitoring, or local vibration sensors. Single physical quantity monitoring can only reflect the local state and is insufficient to comprehensively assess the flow status of powder and the risk of agglomeration throughout the entire discharging channel. Furthermore, existing automated control systems typically rely on empirical parameter settings, such as setting fixed rotational speeds or vibration amplitudes for the discharge mechanism. However, this empirical control method cannot cope with dynamic changes under different material properties and production process conditions, and it also lacks the ability to intelligently analyze the instantaneous discharge state of powder. Meanwhile, the powder discharge process is significantly affected by production process parameters such as reactor temperature and pressure, and there is a complex nonlinear relationship between these parameters and the discharge state.

[0003] This application uses multi-source sensors to collect real-time data on powder discharge status and production parameters, and combines this with a deep learning model to predict flow and agglomeration risks, thereby achieving intelligent monitoring and control of the discharge process, improving discharge stability, reducing agglomeration rate, and increasing production efficiency. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application proposes an intelligent monitoring system and method for material discharge in pharmaceutical production.

[0005] To achieve the above objectives, this application provides the following technical solution:

[0006] A smart monitoring system and method for material discharge in pharmaceutical production includes the following specific steps:

[0007] Obtain the discharge characteristic parameters during the powder discharge process;

[0008] Based on the acquired discharge characteristic parameters, discharge state features are extracted to generate a powder discharge state feature vector;

[0009] Obtain the production parameters of the reactor and conduct a risk assessment of powder discharge based on the reactor production parameters and discharge status characteristics;

[0010] The degree of risk impact is determined based on the risk of powder discharge, and early warning and control measures are implemented.

[0011] Preferably, the process of obtaining the discharge characteristic parameters during powder discharge includes the following specific steps:

[0012] S11. During the powder discharge process, the cumulative discharge weight value and corresponding time information are obtained in real time through the weighing device, and the mechanical vibration signal generated by the discharge channel, including vibration acceleration and vibration amplitude signals, is collected in real time through the vibration sensor of the powder discharge channel to reflect the flow state of the powder in the discharge channel.

[0013] S12. Read the real-time motor current value through the motor drive controller to reflect the degree of mechanical resistance of the powder to the discharge mechanism;

[0014] S13. In the discharge area, the sound signal during the powder discharge process is collected in real time by a sound pickup device.

[0015] Preferably, the step of extracting discharge state features based on the acquired discharge characteristic parameters and generating a powder discharge state feature vector includes the following specific steps:

[0016] S21. Based on the characteristics of the discharge rate change and the degree of fluctuation obtained from the data of the change of powder discharge weight over time, the weight change is calculated by the discharge weight between two adjacent sampling times. The weight change is divided by the corresponding sampling time interval to obtain the weight change slope within the corresponding time period. The weight change slope is used to represent the change of powder discharge weight per unit time. Within a preset time window, multiple weight change slopes are statistically analyzed, and the deviation between each weight change slope and the corresponding average slope within the time window is calculated. The degree of deviation is used as the fluctuation amplitude of weight change. The fluctuation amplitude of weight change is used to represent the stability of the weight change slope within the time window.

[0017] S22. Based on the vibration signal, the dominant frequency characteristics and energy distribution characteristics are obtained. Within a preset time window, the vibration signal is frequency domain transformed by fast Fourier transform. The frequency component with the largest energy proportion is selected as the dominant frequency characteristic of vibration. Within the set frequency range, the energy of each frequency band of the vibration signal is statistically analyzed, and the proportion of signal energy in each frequency band to the total energy is calculated to obtain the vibration energy distribution characteristics.

[0018] S23. Based on the motor current data obtained by continuous sampling, calculate the change in motor current between adjacent sampling times, and use the change as a resistance change feature to reflect the degree of mechanical resistance generated by the powder to the discharge mechanism during the discharge process.

[0019] S24. Perform frequency domain analysis on the acquired sound signal. Based on the frequency range of powder flow behavior, statistically analyze the sound energy in the frequency band as the effective frequency band energy feature. The effective frequency band energy is used to represent the energy of the sound signal in the frequency band related to the collision and friction with the powder during the powder discharge process.

[0020] S25. Generate a powder discharge state feature vector from the weight change slope and fluctuation amplitude, vibration main frequency and energy distribution characteristics, discharge resistance change characteristics, and sound effective frequency band energy characteristics obtained within the same time window.

[0021] Preferably, the step of obtaining the reactor production parameters and conducting a powder discharge risk assessment based on the reactor production parameters and discharge status characteristics includes the following specific steps:

[0022] S31. Normalize each parameter in the powder discharge state feature vector and convert it into a uniform dimensionless numerical range. Then, weight the normalized parameters according to the sensitivity of the corresponding discharge anomalies to obtain the real-time stability evaluation value.

[0023] S32. Acquire the temperature and pressure data recorded in the reactor during the production process. Combine the powder discharge state feature vector with the temperature and pressure data recorded in the reactor during the production process. Preprocess the acquired temperature and pressure time series data, remove obviously abnormal data points, smooth the data, reduce the impact of random noise on the prediction results, and input the data into a pre-trained deep learning neural network model for powder discharge risk assessment. Output discharge flow risk index and agglomeration induction risk index. The discharge flow risk index is used to reflect the risk level of discontinuous flow and discharge rate fluctuation of powder during the discharge process. The agglomeration induction risk index is used to reflect the risk level of bridging and agglomeration structure formed by powder in the discharge channel.

[0024] Preferably, the step of determining the degree of risk impact based on powder discharge risk and conducting early warning and control includes the following specific steps:

[0025] S41. Compare the real-time stability evaluation value with the preset comprehensive early warning threshold. When the real-time stability evaluation value is less than the preset comprehensive early warning threshold, the material discharge process is stable and continues the current material discharge state. When the real-time stability evaluation value is greater than or equal to the preset comprehensive early warning threshold, the material discharge process is unstable and a primary early warning is triggered.

[0026] S42. When in the initial warning state, the discharge flow risk index and agglomeration induction risk index are compared with the preset flow threshold and agglomeration threshold, respectively. When the discharge flow risk index is less than the preset flow threshold, it is judged as normal discharge. When the discharge flow risk index is greater than or equal to the preset flow threshold, it is judged as abnormal discharge flow. The speed of the discharge mechanism drive motor is reduced within the preset range, and the vibration amplitude of the vibration device is increased within the preset range. During the adjustment process, the powder discharge flow risk index is continuously evaluated. When the agglomeration induction risk index is less than the preset agglomeration threshold, it is judged as normal discharge. When the agglomeration induction risk index is greater than or equal to the preset agglomeration threshold, it is judged as abnormal discharge agglomeration. The vibration amplitude of the vibration device is gradually increased within the preset range, while keeping the speed of the discharge mechanism drive motor unchanged. During the adjustment process, the powder agglomeration induction risk index is continuously evaluated.

[0027] The intelligent monitoring system for material discharge in pharmaceutical production is based on the aforementioned intelligent monitoring method for material discharge in pharmaceutical production, and specifically includes:

[0028] The data acquisition module is used to acquire the discharge characteristic parameters during the powder discharge process;

[0029] The powder discharge status extraction module is used to extract discharge status features by acquiring discharge feature parameters and generate a powder discharge status feature vector.

[0030] The powder discharge risk assessment module is used to obtain the reactor production parameters and conduct powder discharge risk assessment based on the reactor production parameters and discharge status characteristics.

[0031] The discharge risk early warning module is used to determine the degree of risk impact based on the powder discharge risk, and to provide early warning and control.

[0032] An electronic device includes: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0033] The processor executes the above-described intelligent monitoring method for drug production output by calling the computer program stored in the memory.

[0034] A computer-readable storage medium is characterized by storing instructions that, when executed on a computer, cause the computer to perform the above-described intelligent monitoring method for material discharge in pharmaceutical production.

[0035] Compared with the prior art, the beneficial effects of this application are:

[0036] This application obtains discharge characteristic parameters during the powder discharge process, extracts discharge state characteristics based on the obtained discharge characteristic parameters, generates a powder discharge state feature vector, obtains reactor production parameters, conducts powder discharge risk assessment based on reactor production parameters and discharge state characteristics, determines the degree of risk impact based on powder discharge risk, and performs early warning and control to improve discharge stability and increase production efficiency. Attached Figure Description

[0037] Figure 1 This is a schematic diagram of the overall process of the intelligent monitoring method for material discharge in pharmaceutical production used in this application.

[0038] Figure 2 This is a flowchart of the material discharge status feature extraction process for this application;

[0039] Figure 3 This is a flowchart of the powder discharge risk assessment for this application;

[0040] Figure 4 This is a schematic diagram of the overall framework of the intelligent monitoring system for material discharge in pharmaceutical production, which is the subject of this application. Detailed Implementation

[0041] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0042] Example 1

[0043] Please see Figure 1 - Figure 3 One embodiment provided in this application is a method for intelligent monitoring of material discharge in pharmaceutical production, which includes the following specific steps:

[0044] Obtain the discharge characteristic parameters during the powder discharge process;

[0045] Based on the acquired discharge characteristic parameters, discharge state features are extracted to generate a powder discharge state feature vector;

[0046] Obtain the production parameters of the reactor and conduct a risk assessment of powder discharge based on the reactor production parameters and discharge status characteristics;

[0047] The degree of risk impact is determined based on the risk of powder discharge, and early warning and control measures are implemented.

[0048] In this embodiment, it should be specifically explained that obtaining the discharge characteristic parameters during the powder discharge process includes the following specific steps:

[0049] S11. During the powder discharge process, the cumulative discharge weight and corresponding time information are obtained in real time through a weighing device. A weighing sensor is installed below the powder discharge container. The weighing sensor outputs the current cumulative discharge weight according to a preset sampling period to reflect the changing trend of the powder discharge amount. The vibration sensor of the powder discharge channel collects the mechanical vibration signal generated by the discharge channel in real time, including vibration acceleration and vibration amplitude signals, to reflect the flow state of the powder in the discharge channel and indicate the changes in vibration characteristics when bridging and agglomeration occur. The mechanical action of the powder on the discharge channel is different when it is in normal flow, bridging, or agglomeration and breaking, resulting in changes in vibration characteristics.

[0050] S12. Read the real-time motor current value through the motor drive controller. When the powder flows poorly in the discharge channel, or when it bridges or clumps, the resistance of the discharge mechanism increases, causing the motor current to change. This is used to reflect the degree of mechanical obstruction of the powder to the discharge mechanism.

[0051] S13. In the discharge area, the sound signal of the powder during the discharge process is collected in real time by the sound pickup device. Different flow states of the powder correspond to different sound characteristics. When the powder is discharged continuously and uniformly, the sound is stable. When the discharge is interrupted or the powder clumps fall off, the sound changes.

[0052] In this embodiment, it should be specifically explained that the extraction of discharge state features based on the acquired discharge characteristic parameters and the generation of powder discharge state feature vectors include the following specific steps:

[0053] S21. Based on the characteristics of the discharge rate change and fluctuation degree obtained from the data of powder discharge weight change over time, the weight change is calculated by the discharge weight between two adjacent sampling times. The weight change is divided by the corresponding sampling time interval to obtain the weight change slope within the corresponding time period. The weight change slope is used to represent the change of powder discharge weight per unit time. Within a preset time window, multiple weight change slopes are statistically analyzed, and the deviation between each weight change slope and the corresponding average slope within the time window is calculated. The deviation degree is used as the fluctuation amplitude of weight change. The fluctuation amplitude of weight change is used to represent the stability of the weight change slope within the time window. The larger the fluctuation amplitude, the more unstable the powder discharge. The length of the time window is set according to the characteristics of the discharged material. The deviation degree is used to characterize the stability of the instantaneous discharge rate during the powder discharge process. The larger the value, the more obvious the powder discharge fluctuation, which is used to characterize the instantaneous discharge rate of the powder.

[0054] S22. Based on the vibration signal, the dominant frequency characteristics and energy distribution characteristics are obtained. Within a preset time window, the vibration signal is frequency-domain converted by fast Fourier transform. The frequency component with the largest energy proportion is selected as the dominant frequency characteristic of vibration. When bridging or agglomeration occurs, the dominant frequency of vibration will shift or weaken. Within the set frequency range, the energy of each frequency band of the vibration signal is statistically analyzed, and the proportion of signal energy in each frequency band to the total energy is calculated to obtain the vibration energy distribution characteristics. The change in energy distribution is used to reflect the process of powder flow from continuous to discontinuous.

[0055] S23. Based on the motor current data obtained by continuous sampling, calculate the change in motor current between adjacent sampling times, and use the change as a resistance change feature to reflect the degree of mechanical resistance generated by the powder to the discharge mechanism during the discharge process. The resistance change feature is used to indicate the short-term load change trend of the discharge mechanism during the discharge process. When the flow of powder is obstructed, the load of the discharge mechanism increases.

[0056] S24. Perform frequency domain analysis on the acquired sound signal. Based on the frequency range of powder flow behavior, statistically analyze the sound energy within the frequency band as the effective frequency band energy characteristic. The effective frequency band energy is used to represent the energy magnitude of the sound signal in the frequency band related to collision and friction with the powder during the powder discharge process.

[0057] In this embodiment, the frequency range of powder flow behavior is determined by analyzing historical powder flow sound signal data. For example, under different powder flow states, such as continuous uniform discharge, bridging formation, and agglomeration and breakup, a large number of discharge sound signals are collected. The collected sound signals are analyzed in the frequency domain by fast Fourier transform, and the energy distribution characteristics of different frequency components are statistically analyzed. Based on the sound energy concentration area under different flow states, the frequency range related to powder collision, friction and vibration is determined. Through multiple experiments, interference noise and irrelevant frequency bands are eliminated to obtain a stable frequency range that can effectively reflect the powder flow state.

[0058] S25. Generate a powder discharge state feature vector from the weight change slope and fluctuation amplitude, vibration main frequency and energy distribution characteristics, discharge resistance change characteristics, and sound effective frequency band energy characteristics obtained within the same time window.

[0059] In this embodiment, it is necessary to specifically explain that obtaining the reactor production parameters and conducting a powder discharge risk assessment based on the reactor production parameters and discharge status characteristics includes the following specific steps:

[0060] S31. Normalize each parameter in the powder discharge state feature vector to convert it into a uniform dimensionless numerical range. Weight the normalized parameters according to the sensitivity of the corresponding discharge anomaly to obtain a real-time stability evaluation value. The sensitivity of the corresponding discharge anomaly is obtained by collecting a large amount of historical data in the powder discharge process, including the slope and fluctuation amplitude of weight change, vibration main frequency and energy distribution, motor current change, and sound effective frequency band energy characteristics. At the same time, it is marked whether a discharge anomaly occurs in each time period. Statistical analysis is performed on each feature parameter and the occurrence of discharge anomalies. For example, the Pearson correlation coefficient and mutual information are calculated to quantify the change amplitude and sensitivity of the feature parameter when the anomaly occurs. All feature parameters are sorted according to the sensitivity index. The sensitivity index is normalized to obtain the weight of each feature parameter, so that the sum of the weights is 1.

[0061] S32. Acquire temperature and pressure data recorded in the reactor during production. Combine the powder discharge state feature vector with the temperature and pressure data recorded in the reactor during production. Preprocess the acquired temperature and pressure time series data, removing obviously abnormal data points and smoothing the data to reduce the impact of random noise on the prediction results. Input the data into a pre-trained deep learning neural network model for powder discharge risk assessment, outputting discharge flow risk indicators and agglomeration induction risk indicators. The discharge flow risk indicator reflects the degree of risk of discontinuous flow and fluctuation in discharge rate during powder discharge. The agglomeration induction risk indicator... The indicators are used to reflect the risk level of powder forming bridging or agglomeration structures in the discharge channel. Among them, the discharge flow risk index and the agglomeration-induced risk index are quantified based on historical powder discharge process data. The specific steps of the deep learning neural network model are as follows: 1. Model construction: The input layer is the powder discharge state feature vector (including weight change slope and fluctuation amplitude, vibration main frequency and energy distribution characteristics, discharge resistance change characteristics, and sound effective frequency band energy characteristics) and the production parameters of reactor temperature and pressure. The first hidden layer is a fully connected layer with 128 neurons. The activation function is ReLU, which introduces nonlinearity and accelerates convergence. Dropout is used for regularization. To prevent overfitting, the second hidden layer is a fully connected layer with 64 neurons, using ReLU activation function and regularization of 0.3. The third hidden layer is a fully connected layer with 32 neurons, using ReLU activation function and regularization of 0.2. The output layer is a fully connected layer with 2 neurons, corresponding to the output flow risk index and agglomeration induction risk index, using Linear activation function to ensure that the predicted values ​​are non-negative continuous values. ReLU can be selected according to the actual situation. 2. Model training configuration: The loss function is mean squared error (MSE), the optimizer is Adam optimizer, the learning rate is set to 0.001, the evaluation metric is mean absolute error (MAE), the batch size is 32, the number of training iterations is 200, and the validation set is divided into training and training sets. 20% of the data is used to monitor model overfitting; 3. Model training and evaluation process: The historically collected powder discharge status characteristics and reactor production parameter data are divided into training set, validation set and test set (e.g. 70% / 15% / 15%). The training set is used to fit the model, and the loss is evaluated on the validation set at the end of each batch. The training and validation loss curves are observed. Both should decrease synchronously and eventually tend to stabilize. If the training loss decreases while the validation loss increases, it indicates overfitting. After training, the model is evaluated on an unseen test set to obtain the model's generalization performance. The trained model is saved. When a new discharge task is completed, the model can be called to predict the discharge flow risk index and agglomeration-induced risk index in real time, providing a basis for online risk assessment and control.

[0062] For example, in this embodiment, the quantification process of the discharge flow risk index includes: performing time series processing on the real-time collected powder discharge weight data, calculating the weight change per unit time through adjacent sampling points to obtain the weight change slope, calculating the average value of multiple weight change slopes within a preset time window, and statistically analyzing the deviation between each slope and the average value to reflect the fluctuation range of the discharge rate, then normalizing the weight change slope and fluctuation range characteristics, and weighting its abnormal sensitivity in combination with historical data to form a discharge rate stability evaluation value, mapping the discharge rate stability evaluation value to the numerical range of the discharge flow risk index, the larger the index value, the more obvious the discharge rate fluctuation and the more unstable the discharge flow, the lower the index value when the powder flow is continuous and uniform, and the rapid increase of the index value when the discharge is intermittent, suddenly stopped or the rate fluctuates drastically;

[0063] For example, in this embodiment, the quantification process of the agglomeration-induced risk index includes: real-time acquisition of vibration signals from the discharge channel using vibration sensors, performing fast Fourier transform on the signals to extract the main frequency characteristics and energy distribution of each frequency band, and analyzing the vibration frequency shift or energy concentration changes through a time window to reflect the trend of powder transitioning from a continuous flow state to a discontinuous flow state. Combined with the current changes collected by the motor drive controller, the short-term load change is calculated to obtain the discharge resistance change characteristics. When the powder bridges or agglomerates, the motor load increases significantly, and the resistance change characteristic value rises. In the sound signals acquired by sound pickup in the discharge area, the effective frequency band energy related to powder collision and friction is statistically analyzed in the frequency domain. The vibration, resistance, and sound characteristics are normalized and weighted according to the weight of each characteristic's sensitivity to agglomeration to obtain the agglomeration value. The agglomeration value is then mapped to the numerical range of the agglomeration-induced risk index.

[0064] In this embodiment, it is necessary to specifically explain that the determination of the degree of risk impact based on the risk of powder discharge, and the implementation of early warning and control, include the following specific steps:

[0065] S41. Compare the real-time stability evaluation value with the preset comprehensive early warning threshold. When the real-time stability evaluation value is less than the preset comprehensive early warning threshold, the material discharge process is stable and continues the current material discharge state. When the real-time stability evaluation value is greater than or equal to the preset comprehensive early warning threshold, the material discharge process is unstable and a primary early warning is triggered.

[0066] S42. When in the initial warning state, the discharge flow risk index and agglomeration induction risk index are compared with the preset flow threshold and agglomeration threshold, respectively. When the discharge flow risk index is less than the preset flow threshold, it is judged as normal discharge. When the discharge flow risk index is greater than or equal to the preset flow threshold, it is judged as abnormal discharge flow. The speed of the discharge mechanism drive motor is reduced within the preset range, and the vibration amplitude of the vibration device is increased within the preset range. During the adjustment process, the powder discharge flow risk index is continuously evaluated. When the agglomeration induction risk index is less than the preset agglomeration threshold, it is judged as normal discharge. When the agglomeration induction risk index is greater than or equal to the preset agglomeration threshold, it is judged as abnormal discharge agglomeration. The vibration amplitude of the vibration device is gradually increased within the preset range, while keeping the speed of the discharge mechanism drive motor unchanged. During the adjustment process, the powder agglomeration induction risk index is continuously evaluated.

[0067] The advantages of this embodiment compared to the prior art are:

[0068] This application obtains discharge characteristic parameters during the powder discharge process, extracts discharge state characteristics based on the obtained discharge characteristic parameters, generates a powder discharge state feature vector, obtains reactor production parameters, conducts powder discharge risk assessment based on reactor production parameters and discharge state characteristics, determines the degree of risk impact based on powder discharge risk, and performs early warning and control to improve discharge stability and increase production efficiency.

[0069] Example 2

[0070] like Figure 4 As shown, the intelligent monitoring system for material discharge in pharmaceutical production is based on the aforementioned intelligent monitoring method for material discharge in pharmaceutical production. Specifically, it includes a data acquisition module, a powder discharge status extraction module, a powder discharge risk assessment module, and a discharge risk early warning module. The data acquisition module acquires discharge characteristic parameters during the powder discharge process; the powder discharge status extraction module extracts discharge status features from the acquired discharge characteristic parameters to generate a powder discharge status feature vector; the powder discharge risk assessment module acquires reactor production parameters and assesses powder discharge risk based on these parameters and discharge status features; and the discharge risk early warning module determines the degree of risk impact based on powder discharge risk and provides early warning and control.

[0071] Example 3

[0072] This embodiment provides an electronic device, including: a processor and a memory, wherein the memory stores a computer program that can be called by the processor;

[0073] The processor executes the aforementioned intelligent monitoring method for drug production output by calling computer programs stored in memory.

[0074] The electronic device can vary considerably depending on its configuration and performance. It may include one or more Central Processing Units (CPUs) and one or more memories, wherein the memory stores at least one computer program, which is loaded and executed by the processor to implement the intelligent monitoring method for pharmaceutical production discharge provided in the above-described embodiment. The electronic device may also include other components for implementing its functions; for example, it may have wired or wireless network interfaces and input / output interfaces for data input and output. Details will not be elaborated upon in this embodiment.

[0075] Example 4

[0076] This embodiment proposes a computer-readable storage medium on which an erasable and rewritable computer program is stored.

[0077] When the computer program runs on the computer device, it causes the computer device to execute the above-mentioned intelligent monitoring method for the discharge of pharmaceuticals.

[0078] For example, computer-readable storage media can be read-only memory (ROM), random access memory (RAM), compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, and optical data storage devices.

[0079] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions according to the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired network and / or wireless network. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

Claims

1. A method for intelligent monitoring of material discharge in pharmaceutical production, characterized in that, It includes the following specific steps: Obtain the discharge characteristic parameters during the powder discharge process; Based on the acquired discharge characteristic parameters, discharge state features are extracted, including weight change slope and fluctuation amplitude, vibration main frequency and energy distribution features, discharge resistance change features and sound effective frequency band energy features, generating a powder discharge state feature vector. Obtain the reactor production parameters and conduct a powder discharge risk assessment based on the reactor production parameters and discharge status characteristics, including the following specific steps: Each parameter in the feature vector of powder discharge state is normalized, and the normalized parameters are weighted and combined to obtain the real-time stability evaluation value. The temperature and pressure data recorded in the reactor during the production process are obtained. The powder discharge state feature vector and the temperature and pressure data recorded in the reactor during the production process are input into a pre-trained deep learning neural network model to conduct powder discharge risk assessment and output discharge flow risk index and agglomeration induced risk index. The degree of risk impact is determined based on the risk of powder discharge, and early warning and control measures are implemented.

2. The intelligent monitoring method for material discharge in pharmaceutical production as described in claim 1, characterized in that, The specific steps for obtaining the discharge characteristic parameters during the powder discharge process include the following: S11. During the powder discharge process, the cumulative discharge weight value and corresponding time information are obtained in real time through the weighing device, and the mechanical vibration signal generated by the discharge channel is collected in real time through the vibration sensor of the powder discharge channel. S12. Read the real-time motor current value through the motor drive controller; S13. In the discharge area, the sound signal during the powder discharge process is collected in real time by a sound pickup device.

3. The intelligent monitoring method for material discharge in pharmaceutical production as described in claim 2, characterized in that, The process of extracting discharge state features based on the acquired discharge characteristic parameters and generating a powder discharge state feature vector includes the following specific steps: S21. Calculate the weight change by the weight of the material discharged between two adjacent sampling times, divide the weight change by the corresponding sampling time interval to obtain the weight change slope within the corresponding time period, statistically analyze multiple weight change slopes within a preset time window, and calculate the degree of deviation between each weight change slope within the time window and the corresponding average slope, and use the degree of deviation as the fluctuation range of weight change. S22. Within a preset time window, the vibration signal is frequency-domain converted by fast Fourier transform. The frequency component with the largest energy proportion is selected as the vibration main frequency feature. Within the set frequency range, the energy of each frequency band of the vibration signal is statistically analyzed, and the proportion of signal energy in each frequency band to the total energy is calculated to obtain the vibration energy distribution characteristics. S23. Based on the motor current data obtained by continuous sampling, calculate the change in motor current between adjacent sampling times, and use the change as a characteristic of resistance change. S24. Perform frequency domain conversion on the acquired sound signal, and based on the frequency range of powder flow behavior, statistically analyze the sound energy within the frequency band as the effective frequency band energy feature of the sound. S25. Generate a powder discharge state feature vector from the weight change slope and fluctuation amplitude, vibration main frequency and energy distribution characteristics, discharge resistance change characteristics, and sound effective frequency band energy characteristics obtained within the same time window.

4. The intelligent monitoring method for material discharge in pharmaceutical production as described in claim 3, characterized in that, The process of determining the degree of risk impact based on powder discharge risk, and then conducting early warning and control, includes the following specific steps: S41. Compare the real-time stability evaluation value with the preset comprehensive early warning threshold. When the real-time stability evaluation value is less than the preset comprehensive early warning threshold, the material discharge process is stable and continues the current material discharge state. When the real-time stability evaluation value is greater than or equal to the preset comprehensive early warning threshold, the material discharge process is unstable and a primary early warning is triggered. S42. When in the initial warning state, the discharge flow risk index and agglomeration induction risk index are compared with the preset flow threshold and agglomeration threshold, respectively. When the discharge flow risk index is less than the preset flow threshold, it is judged as normal discharge. When the discharge flow risk index is greater than or equal to the preset flow threshold, it is judged as abnormal discharge flow. The speed of the discharge mechanism drive motor is reduced within the preset range, and the vibration amplitude of the vibration device is increased within the preset range. During the adjustment process, the powder discharge flow risk index is continuously evaluated. When the agglomeration induction risk index is less than the preset agglomeration threshold, it is judged as normal discharge. When the agglomeration induction risk index is greater than or equal to the preset agglomeration threshold, it is judged as abnormal discharge agglomeration. The vibration amplitude of the vibration device is gradually increased within the preset range, while keeping the speed of the discharge mechanism drive motor unchanged. During the adjustment process, the powder agglomeration induction risk index is continuously evaluated.

5. An intelligent monitoring system for material discharge in pharmaceutical production, implemented based on the intelligent monitoring method for material discharge in pharmaceutical production as described in any one of claims 1-4, characterized in that, Specifically, it includes: The data acquisition module is used to acquire the discharge characteristic parameters during the powder discharge process; The powder discharge status extraction module is used to extract discharge status features by acquiring discharge feature parameters and generate a powder discharge status feature vector. The powder discharge risk assessment module is used to obtain the reactor production parameters and conduct powder discharge risk assessment based on the reactor production parameters and discharge status characteristics. The discharge risk early warning module is used to determine the degree of risk impact based on the powder discharge risk, and to provide early warning and control.

6. An electronic device, comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; The processor is characterized in that it executes the intelligent monitoring method for material discharge in pharmaceutical production as described in any one of claims 1-4 by calling a computer program stored in the memory.

7. A computer-readable storage medium, characterized in that, The device stores instructions that, when executed on a computer, cause the computer to perform the intelligent monitoring method for material discharge in pharmaceutical production as described in any one of claims 1-4.

Images