Discharging amount prediction method, concentration prediction method and system for continuous powder feeding

Abstract

The invention discloses a discharging amount prediction method, concentration prediction method and system for continuous powder feeding. The discharging amount prediction method comprises the steps that the real-time feeding amount of a mixing machine in unit time is obtained; inputting the real-time feeding amount into a preset discharging amount real-time prediction model, and predicting to obtain the real-time discharging amount of the mixing machine; wherein RTD sample data is stored in the preset discharging amount real-time prediction model; the RTD sample data is updated in real time according to the real-time feeding amount. The method has the advantages of high prediction precision, low hardware requirement, high compatibility and the like.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of food and medicine packaging, and in particular to a method for predicting the discharge amount of powder continuous feeding, a method for predicting the concentration and a system thereof. Background Art

[0002] In the powder production process, the prediction of powder concentration is a crucial link. Traditional powder concentration measurement methods are often complicated to operate and have low prediction accuracy, which cannot meet the high efficiency and precision requirements of modern production. Therefore, it is particularly important to develop a simple, fast and accurate powder concentration measurement method.

[0003] RTD (Residence Time Distribution) is a method to describe the residence time distribution of particles in a mixture. The residence time is a key indicator to describe the uniformity of mixing, and its basic mathematical framework is suitable for all continuous flow systems.

[0004] As powder enters the mixer at a steady-state rate, some particles experience greater convection in the forward axial direction, while some experience an anomalous push backward or become trapped in the mixer dead zone for a period of time. Therefore, while most particles spend a period of time in the mixer close to some average value (defined as the “mean residence time”), some particles may leave the mixer quickly while others spend an anomalous amount of time in the mixer. Therefore, for a representative set of particles entering the mixer, there is a residence time distribution, or RTD, at steady state.

[0005] like Figure 1 As shown, the initially sharp pulse broadens as it travels through the length of the mixer. Some particles leave the mixer quickly, and as the tracer travels through the mixer tube, the sharp pulse broadens due to axial convection while some particles take a long time to leave the system. The time that most particles pass through is close to the average, which can be calculated by calculating the average of the distribution or by dividing the mass retention of the mixer by the operating mass flow rate. In the absence of dead space, these two methods of calculating the average residence time give similar results. Specifically, the RTD function E(t) is defined as:

[0006]

[0007] E(t) is the discharge probability at the current moment; C(t) is the material concentration at the current moment;

[0008] The predicted discharge and average residence time (or average time the particles are in the mixer) are given by:

[0009]

[0010] V inis the feed amount at the current moment; V out is the predicted discharge volume; τ is the average residence time.

[0011] Finally, the mean centered variance (MCV), or a measure of the width of the RTD profile and its noise filtering capabilities, is calculated by:

[0012]

[0013] is the central variance of the experimental sample.

[0014] At present, the mainstream algorithm in the industry is the calculation method based on the RTD (Residence Time Distribution) model, which has several obvious shortcomings:

[0015] 1. The algorithm is too complex, the amount of calculation is large, and the implementation is difficult, so it cannot be directly calculated through PLC.

[0016] 2. Each material needs to be measured through a large number of experiments using tracers. The acquisition of sample curves has high requirements and high requirements for operators.

[0017] 3. RDT itself is a probability distribution with a certain probability. It has too high requirements on material properties and operating conditions. Once a certain factor changes, it will affect the shape of the operating RTD curve. Summary of the invention

[0018] In view of the technical problems existing in the prior art, the present invention provides a method for predicting the discharge amount of continuous powder feeding with high prediction accuracy, a method for predicting the concentration and a system.

[0019] In order to solve the above technical problems, the technical solution proposed by the present invention is:

[0020] A method for predicting the discharge amount of powder continuous feeding comprises the steps of:

[0021] Get the real-time feed amount per unit time of the mixer;

[0022] The real-time feed amount is input into a preset real-time discharge amount prediction model to predict the real-time discharge amount of the mixer;

[0023] The preset real-time prediction model for discharge volume stores RTD sample data; the RTD sample data is updated in real time according to the real-time feed volume.

[0024] Preferably, the specific process of updating the RTD sample data in real time according to the real-time feed amount is as follows:

[0025] Establish a matrix M to store RTD sample data;

[0026] The matrix M is moved down and right according to the real-time feed amount; the matrix M is moved down to ensure that the data in the matrix M matches the actual position of the material; the matrix is ​​moved right to ensure that the data in the matrix can accurately reflect the spatial distribution of the material;

[0027] Then, according to the real-time feed amount and the RTD sample data after being moved in the matrix M, new RTD sample data are obtained.

[0028] Preferably, the length of the data storage area is 300.

[0029] Preferably, the spatial size of the matrix M is 300*300.

[0030] Preferably, the RTD sample data is data at the material outlet position.

[0031] The present invention also discloses a concentration prediction method for continuous powder feeding, which is used to obtain the real-time discharge amount of at least two materials per unit time of a mixer according to the discharge amount prediction method for continuous powder feeding as described above, and then calculate the real-time concentration according to the real-time discharge amount of the two materials.

[0032] Preferably, after calculating the concentration of a material at any location, the residence time is calculated to correct the concentration.

[0033] The present invention further discloses a computer program product, comprising a computer program, wherein the computer program executes the steps of the method described above when executed by a processor.

[0034] The present invention also discloses a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described above are executed.

[0035] The present invention also discloses a discharge amount prediction system for continuous powder feeding, comprising a memory and a processor connected to each other, wherein a computer program is stored in the memory, and when the computer program is run by the processor, the steps of the method described above are executed.

[0036] Compared with the prior art, the advantages of the present invention are:

[0037] The present invention creates a data model (matrix M) that is identical to the actual process and updates the data model in real time, thereby accurately predicting the material outlet concentration through the residence time distribution (RTD) curve of the material and the material inlet flow rate at the corresponding time.

[0038] The present invention utilizes a method for calculating the real-time flow rate at the outlet of a continuous flow system by optimizing the RTD model, thereby optimizing the algorithm's requirements for hardware, having lower costs and stronger compatibility; the concentration of materials at any position in the system can be predicted more clearly and intuitively; the optimization algorithm can get rid of dependence on a large amount of experimental data, making calculations simpler and the boundaries between discarding and recovering clearer.

[0039] The method of the present invention optimizes the RTD calculation process and can be directly implemented through PLC program, single-chip microcomputer, industrial computer script or other controllers. It has strong versatility, can realize the overall modeling of complex systems, and can predict the material distribution of the system at any position. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a curve diagram of the existing RTD function E(t).

[0041] Figure 2 It is a structural schematic diagram of the data storage area in the present invention.

[0042] Figure 3 It is a schematic diagram of the state of the matrix M in the present invention moving downward.

[0043] Figure 4 It is a schematic diagram of the state of the matrix M in the present invention being shifted right.

[0044] Figure 5 It is a schematic diagram of the update state of the matrix M in the present invention.

[0045] Figure 6 It is a schematic diagram of the integration of the matrix M in the present invention.

[0046] Figure 7 This is the existing RTD distribution curve.

[0047] Figure 8 It is a comparison diagram of the optimization method of the present invention and the existing RTD curve.

[0048] Fig. 9 It is a comparison diagram of the optimization algorithm of the present invention and the weight loss flow curve.

[0049] Fig.10 This is a comparison chart between the existing RTD algorithm and the loss-in-weight flow curve.

[0050] Fig.11 It is a comparison chart of the optimization algorithm of the present invention, the existing RTD algorithm and the weight loss flow curve.

[0051] Fig.12 It is a deviation analysis diagram between the optimization algorithm of the present invention and the existing RTD algorithm.

[0052] Fig.13The present invention is a flow chart of the method for predicting the discharge amount of continuous powder feeding in an embodiment. DETAILED DESCRIPTION

[0053] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments.

[0054] like Fig.13 As shown, the method for predicting the discharge amount of powder continuous feeding according to an embodiment of the present invention comprises the steps of:

[0055] Get the real-time feed amount per unit time of the mixer;

[0056] The real-time feed amount is input into a preset real-time discharge amount prediction model to predict the real-time discharge amount of the mixer;

[0057] The preset real-time prediction model for discharge volume stores RTD sample data; the RTD sample data is updated in real time according to the real-time feed volume.

[0058] Specifically, the specific process of real-time updating of RTD sample data according to the real-time feed amount is as follows:

[0059] Establish a matrix M to store RTD sample data;

[0060] The matrix M is moved down and right according to the real-time feed amount; the matrix M is moved down to ensure that the data in the matrix M matches the actual position of the material; the matrix is ​​moved right to ensure that the data in the matrix can accurately reflect the spatial distribution of the material;

[0061] Then, according to the real-time feed amount and the RTD sample data after being moved in the matrix M, new RTD sample data are obtained.

[0062] An embodiment of the present invention also discloses a concentration prediction method for continuous powder feeding, which is used to obtain the real-time discharge amount of at least two materials per unit time of a mixer according to the discharge amount prediction method for continuous powder feeding as described above, and then calculate the real-time concentration based on the real-time discharge amount of the two materials.

[0063] The embodiment of the present invention further discloses a computer program product, including a computer program, which executes the steps of the above method when executed by a processor. The embodiment of the present invention further discloses a computer-readable storage medium, which stores a computer program, which executes the steps of the above method when executed by a processor. The embodiment of the present invention further discloses a discharge amount prediction system for continuous powder feeding, including a memory and a processor connected to each other, the memory stores a computer program, and the computer program executes the steps of the above method when executed by the processor. The product, medium and system of the present invention correspond to the above method and also have the advantages described in the above method.

[0064] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0065] 1) Establish a data storage area for storing RTD sample data;

[0066] The RTD sample data collected over time is stored in a data structure, usually an array or matrix, of sufficient length or size to hold the data for the entire time period of interest.

[0067] According to the regulation "Continuous Manufacturing of APIs and Drug Products Q13", the maximum residence time of materials is 300S. Therefore, in order to meet the RDT distribution of all materials, a data storage area with a length of 300 is needed to store the RTD sample data E(t), such as Figure 2 The above data are obtained through tracer experiments at 20 sample points. These sample points are averaged or interpolated in time to ensure uniform distribution of data and accurately reflect the RTD distribution of the entire system at different times.

[0068] 2) Calculate the real-time feed rate per second;

[0069] Real-time feed rate is one of the key parameters of the system's dynamic characteristics. Real-time data is obtained from measuring devices (such as flow meters) and converted into feed rate per second. The feed rate directly affects the concentration and distribution of materials in the system;

[0070] Specifically, the real-time feed volume V in (t) = V / 3600*1000(g);

[0071] 3) Establish a matrix M for storing temporary calculation data;

[0072] like Figure 3As shown, matrix M is used to store temporary calculation data, and its size (300x300) reflects the number of independent positions and times that can be measured in the system.

[0073] The matrix is ​​refreshed once a second, that is, the data in the matrix is ​​updated once a second. This step is performed simultaneously with the calculation of the real-time feed amount to ensure that the data in the matrix is ​​up to date.

[0074] 4) Matrix downshift:

[0075] Over time, the position of materials in the system will change. Figure 3 As shown, the matrix downshift operation (i.e., the data in row n is moved to row n+1) reflects this change, ensuring that the data in the matrix matches the actual location of the material. The first row of data is cleared, waiting for the latest data to be stored;

[0076] This step is the process of dynamically updating the matrix, which allows the system to move forward in time while maintaining the integrity of the data in the matrix.

[0077] 5) Matrix right shift:

[0078] Similar to the matrix downshift, the matrix rightshift operation reflects the lateral movement of materials in the system, such as Figure 4 This is because the material will continuously reach new locations during the flow process. This step ensures that the data in the matrix can accurately reflect the spatial distribution of the material.

[0079] 6) Update RTD distribution data:

[0080] Based on the current real-time feed rate and other known parameters, the latest RTD distribution data is calculated. These data will be used to update the first row of the matrix M, as shown in Figure 5 shown.

[0081] This step is data-driven and relies on real-time feed volume and known RTD data. These data are used to calculate the concentration of the material at different locations and update the matrix M.

[0082] 7) Matrix integration:

[0083] like Figure 6 As shown, the data in the matrix is ​​used to calculate the concentration of the material at any location. This usually involves solving partial differential equations to model the flow and diffusion of the material in the system.

[0084] Matrix integration is a key step in calculating material concentration distribution, which converts discrete data in the matrix into continuous concentration distribution.

[0085] Since the E(t) sample is obtained at the material outlet, the concentration at the material outlet is the most accurate;

[0086] 8) If the residence time of the material sample is n seconds, that is, Vout = Vout[n]*3600 / 1000, Vout is the real-time outlet flow of the current device.

[0087] The present invention creates a data model (matrix M) that is identical to the actual process and updates the data model in real time, thereby accurately predicting the material outlet concentration through the residence time distribution (RTD) curve of the material and the material inlet flow rate at the corresponding time.

[0088] The present invention utilizes a method for calculating the real-time flow rate at the outlet of a continuous flow system by optimizing the RTD model, thereby optimizing the algorithm's requirements for hardware, having lower costs and stronger compatibility; the concentration of materials at any position in the system can be predicted more clearly and intuitively; the optimization algorithm can get rid of dependence on a large amount of experimental data, making calculations simpler and the boundaries between discarding and recovering clearer.

[0089] The method of the present invention optimizes the RTD calculation process and can be directly implemented through PLC program, single-chip microcomputer, industrial computer script or other controllers. It has strong versatility, can realize the overall modeling of complex systems, and can predict the material distribution of the system at any position.

[0090] The above method mainly uses the prediction of mass flow rate of continuous flow system and the kicking action, realizes the concentration prediction and qualification judgment of the continuous flow system, improves the product qualification rate and reduces production costs; when the material shows an unqualified trend, the optimization algorithm can amplify the unqualified signal and send out the kicking signal in advance.

[0091] like Figure 7 The curve shown is obtained through tracer experiments. This curve is related to the material properties and the residence time in the equipment. When a certain factor changes, the curve will deform or drift. Therefore, some fault-tolerant operations need to be performed on the prediction results when making qualification judgments.

[0092] The longer the residence time, the flatter the curve will be. The ideal state is that the material concentration is uniform at every moment.

[0093] Based on the above two points, assuming that the material enters the equipment uniformly, the corresponding curve is as follows Figure 8 shown.

[0094] Simulate a set of data, the target flow rate is 100kg / H, the fluctuation is + / -10% (90kg / H~100kg / H) and the effect after the material enters the mixer is as follows: Figure 9-11 shown.

[0095] After comparison, the prediction results of the two algorithms can be stabilized within 1%, such as Fig.12shown.

[0096] The simulation data of this group with + / -10% deviation and + / -5% were simulated respectively, and the results were as follows:

[0097]

[0098]

[0099] After experimental verification, when the feed fluctuation is + / -10%, the average deviation between the optimized algorithm of the present invention and the RTD algorithm result is 0.44%, and when the feed fluctuation is + / -5%, the average deviation between the optimized algorithm and the RTD algorithm result is 0.02%. At the same time, when the feed fluctuates, the optimized algorithm can amplify the fluctuation signal and turn on the kick signal in advance. When the fluctuation disappears, the kick signal is closed with a delay. This effect is exactly the ideal effect required for production.

[0100] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A method for predicting the discharge amount of powder continuous feeding, characterized in that: Includes steps: Get the real-time feed amount per unit time of the mixer; The real-time feed amount is input into a preset real-time discharge amount prediction model to predict the real-time discharge amount of the mixer; The preset real-time prediction model for discharge volume stores RTD sample data; the RTD sample data is updated in real time according to the real-time feed volume.

2. The method for predicting the discharge amount of powder continuous feeding according to claim 1, characterized in that: The specific process of real-time updating of the RTD sample data according to the real-time feed amount is as follows: Establish a matrix M to store RTD sample data; The matrix M is moved down and right according to the real-time feed amount; the matrix M is moved down to ensure that the data in the matrix M matches the actual position of the material; the matrix is ​​moved right to ensure that the data in the matrix can accurately reflect the spatial distribution of the material; Then, according to the real-time feed amount and the RTD sample data after being moved in the matrix M, new RTD sample data are obtained.

3. The method for predicting the discharge amount of powder continuous feeding according to claim 2, characterized in that: The length of the data storage area is 300.

4. The method for predicting the discharge amount of powder continuous feeding according to claim 1, 2 or 3, characterized in that: The spatial size of matrix M is 300*300.

5. The method for predicting the discharge amount of powder continuous feeding according to claim 1, 2 or 3, characterized in that: The RTD sample data is the data at the material outlet position.

6. A method for predicting the concentration of powder continuous feeding, characterized in that: Used to obtain the real-time discharge amount of at least two materials per unit time of a mixer according to the discharge amount prediction method for continuous powder feeding as described in any one of claims 1-5, and then calculate the real-time concentration based on the real-time discharge amount of the two materials.

7. The method for predicting the concentration of powder continuous feeding according to claim 6, characterized in that: After calculating the concentration of a material at any location, the residence time is calculated to correct for the concentration.

8. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are performed.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the computer program performs the steps of the method according to any one of claims 1 to 5.

10. A system for predicting the discharge amount of powder continuous feeding, comprising a memory and a processor connected to each other, wherein a computer program is stored in the memory, characterized in that: When the computer program is executed by a processor, the computer program performs the steps of the method according to any one of claims 1 to 5.

Images