Method and apparatus for adaptive control of intermittent spray-drying in a spray fluidized bed

Abstract

The application discloses a spray fluidized bed intermittent spraying-drying self-adaptive control method and device. The method comprises the following steps: acquiring experimental data of the intermittent spraying-drying process of the spray fluidized bed under different working conditions; constructing a prediction model reflecting the data relationship between the wall hanging mass and the process parameters based on the experimental data; constructing an empirical model reflecting the relationship between the drying waiting time and the RMS recovery characteristics; determining the critical wall hanging mass of the to-be-controlled section; calculating the single allowed spraying time based on the critical wall hanging mass and the prediction model; and calculating the predicted drying waiting time based on the single allowed spraying time, combining the empirical model, and determining the sound signal sampling time. The self-adaptive control method and device are used for adaptively determining the spraying time and the drying recovery state sampling time in the intermittent spraying-drying process of the spray fluidized bed according to different working condition parameters, thereby improving the reliability of the subsequent acoustic monitoring results.

Description

Technical Field

[0001] This application belongs to the fields of fluidized bed spray technology, acoustic signal processing and computer-aided design technology, specifically relating to an adaptive control method and device for intermittent spray-drying in a spray fluidized bed. Background Technology

[0002] Spray fluidized bed granulation is widely used in chemical, pharmaceutical, food, particle coating, and powder modification processes. Its basic principle is to suspend particles in a fluidized state under the action of fluidizing gas, while simultaneously spraying binder droplets into the bed through nozzles. After wetting, the particles collide, agglomerate, and dry, ultimately forming larger agglomerates. This process combines multiple behaviors such as gas-solid fluidization, droplet spraying, particle wetting, agglomeration growth, drying and solidification, and wall adhesion. The operating conditions change rapidly, and operational stability has a significant impact on product quality and production safety.

[0003] Traditional spray fluidized beds typically employ continuous spraying. While this method is simple to operate, it is prone to problems such as excessive liquid accumulation in localized areas during actual operation. Especially when the binder viscosity is high or the spray rate is high, the droplets sprayed into the bed cannot be uniformly captured and dried by the particles in a timely manner. Some wet particles and binder adhere to the reactor wall, gradually forming a wall-mounted layer. As the wall-mounted layer increases in mass, it absorbs and shields the acoustic signals generated by particle collisions, causing a shift in acoustic emission monitoring results. For systems that rely on acoustic emission signals for flow pattern identification, particle size determination, or instability early warning, wall-mounted adhesion significantly reduces signal reliability and can even lead to misjudgments.

[0004] Currently, the main methods used in industrial production to address the issue of wall adhesion in spray granulation include reducing the spray rate, increasing the fluidizing gas velocity, extending the drying time, periodically stopping the machine for cleaning, or adjusting the spray cycle based on manual experience. While these methods can alleviate wall adhesion to some extent, they still have significant shortcomings, as detailed below: First, existing methods rely heavily on manual experience and lack quantifiable formulas for calculating spray time. Operators typically judge when to stop spraying based on visual observation, pressure drop changes, or historical experience, but these methods cannot clearly specify the permissible spray time for different viscosities, spray rates, and bed sizes.

[0005] Second, traditional threshold control methods typically only trigger an alarm or stop spraying when the monitored index exceeds a certain fixed value, which is essentially a post-event judgment. For spray fluidized beds, the formation of wall adhesion quality is related to factors such as bed size, bed material quantity, spray rate, binder viscosity, and fluidizing gas velocity, making it difficult for a single threshold to adapt to multiple operating conditions.

[0006] Third, existing acoustic emission acquisition methods often neglect the signal differences between the spraying and drying sections. In the spraying section, droplet impact, wet particle adhesion, and local liquid film significantly alter the collision acoustic signal. If the acoustic signal is directly acquired during continuous spraying, the signal will simultaneously include the effects of particle collision, droplet impact, liquid film damping, and wall adhesion, leading to unstable subsequent analysis results.

[0007] Fourth, drying time is usually specified by fixed process parameters, lacking feedback on the actual bed recovery state. Even at the same spray rate and spray time, the degree of wetting, fluidization state, and wall residue may differ in different cycles. A fixed drying time cannot guarantee that the bed is in a relatively stable and comparable state each time an acoustic signal is collected.

[0008] Therefore, it is necessary to establish a method that can adaptively determine the spraying time based on operating parameters and determine the sampling time based on the drying recovery state, so that the spray fluidized bed can obtain a more stable and reliable acoustic emission signal while reducing wall adhesion. Summary of the Invention

[0009] This application provides an adaptive control method and apparatus for intermittent spray-drying in a spray fluidized bed, which is used to adaptively determine the spraying time and the sampling time for the drying recovery state in the intermittent spray-drying process of the spray fluidized bed according to different operating parameters.

[0010] To achieve the above objectives, this application adopts the following technical solution: The first aspect provides an adaptive control method for intermittent spray-drying in a spray fluidized bed, including: Experimental data on the intermittent spray-drying process of a spray fluidized bed under different operating conditions were obtained; the experimental data included process parameter data and wall adhesion mass. Based on the experimental data, a predictive model reflecting the relationship between wall adhesion quality and process parameter data is constructed. An empirical model reflecting the relationship between drying waiting time and RMS recovery characteristics was constructed. Determine the critical wall-hanging quality of the section to be controlled; Based on the critical wall adhesion mass and prediction model, the allowable spraying time for a single spraying cycle is calculated. Based on the single allowable spray time and the empirical model, the predicted drying waiting time is calculated, and the sampling time of the acoustic signal is determined.

[0011] In one possible implementation, the process parameters include binder viscosity, spray rate, bed dimensions, bed height, and fluidizing gas velocity.

[0012] In one possible implementation, the formula for the prediction model is as follows:

[0013] In the formula, Y w C is the dimensionless predicted value of the wall-mounted mass. w For wall-mounted quality coefficient; Oh L The Ohnesorge number is the number of the adhesive droplet, used to characterize the relative interaction between adhesive viscosity, surface tension, and droplet characteristic size; Π L Π is a dimensionless liquid load parameter used to characterize the amount of liquid injected per unit bed mass; U The fluidization intensity parameter is used to characterize the ratio between the apparent fluidization velocity and the minimum fluidization velocity; Oh0, Π L0 Π U0 Oh L Π L Π U The reference values ​​are a, b, and c, which are physical constraint indices.

[0014] In one possible implementation, the formula for the empirical model is as follows:

[0015] In the formula, The drying time is dimensionless. This is the drying waiting time. This corresponds to the spray duration of the cycle. This is the drying time coefficient. Indicates the decrease in spray impact during the i-th cycle, α, η, ε is a non-negative exponent, and ε is a minimal quantity to prevent the denominator from being zero. This represents the drying recovery coefficient for the (i-1)th cycle. This indicates the risk of a decrease in the later stages of drying in the i-th cycle.

[0016] In one possible implementation, the formula for calculating the single allowable spray time is as follows:

[0017] In the formula, This refers to the permissible spray time per application. For a single permissible liquid load, , For the dimensionless wall-mounted mass allowed in a single instance, , The critical wall-mounting mass; For bed equivalent quality, This represents the mass flow rate of the sprayed liquid.

[0018] In one possible implementation, the step of calculating the predicted drying waiting time based on the single allowable spray time and in conjunction with the empirical model, and determining the acoustic signal sampling time, includes: The single allowable spray time is taken as the spray duration of the section to be controlled; The predicted drying waiting time is calculated using the empirical model described above; the calculation formula is as follows: ; Predict the drying waiting time for the i-th cycle. The duration of the spray in the i-th cycle; The moment obtained by adding the predicted drying waiting time and safety delay time to the end of the spraying time is determined as the sampling time of the acoustic signal.

[0019] Secondly, an adaptive control device for intermittent spray-drying in a spray fluidized bed is provided, including: The acquisition module is used to acquire experimental data of the intermittent spray-drying process of the spray fluidized bed under different operating conditions; the experimental data includes process parameter data and wall adhesion mass; The first construction module is used to construct a predictive model based on the experimental data, reflecting the relationship between the wall adhesion quality and process parameter data. The second building module is used to construct an empirical model that reflects the relationship between drying waiting time and RMS recovery characteristics; The determination module is used to determine the critical wall adhesion quality of the section to be controlled; The first calculation module is used to calculate the allowable spraying time for a single spraying operation based on the critical wall-mounted mass and the prediction model. The second calculation module is used to calculate the predicted drying waiting time based on the single allowable spray time and the empirical model, and to determine the sampling time of the acoustic signal.

[0020] Thirdly, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the spray fluidized bed intermittent spray-drying adaptive control method as described in the first aspect.

[0021] Fourthly, a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the spray fluidized bed intermittent spray-drying adaptive control method as described in the first aspect.

[0022] Fifthly, a computer program product includes a computer program that, when executed by a processor, implements the spray fluidized bed intermittent spray-drying adaptive control method as described in the first aspect.

[0023] The beneficial effects provided by this application are as follows: (1) The method provided in this application transforms continuous spray control into calculable intermittent spray control. The method provided in this application establishes a dimensionless formula for predicting wall adhesion quality, incorporating binder viscosity, spray load, fluidization intensity, and bed dimensions into the same calculation framework. Compared with methods that rely on experience to set spray time, this method can deduce the allowable spray time per cycle based on different operating conditions, transforming spray control from empirical judgment to quantitative calculation.

[0024] (2) The method provided in this application embodiment can reduce the interference of wall-mounted material on acoustic emission monitoring. The method provided in this application embodiment adopts an intermittent spray-drying operation, which weakens or peels off the wet material layer on the wall during the drying section, avoiding continuous thickening of the wall-mounted layer caused by continuous spraying. Since the acoustic emission signal is mainly collected after drying recovery, the influence of droplet impact, liquid film damping and the shielding of the wet material layer on the acoustic signal is reduced, thereby improving the reliability of subsequent acoustic monitoring results.

[0025] (3) The method provided in this application improves adaptability under different spraying conditions. The method provided in this application uses dimensionless parameters to represent the wall adhesion quality, which can take into account factors such as bed size, spray volume, viscosity, and fluidizing gas velocity. Compared with fixed spray time or single threshold control, this method is more suitable for spray fluidized bed control under different viscosities, different spray rates, and different equipment sizes.

[0026] (4) The method provided in this application provides a method for determining the drying sampling time. The method provided in this application constructs an empirical formula for the drying waiting time using RMS recovery features, and sets the sampling time to be 1 second after the predicted drying time after the spraying ends. This method can avoid prematurely collecting unstable wet signals, making the acoustic emission data closer to the stable dry fluidized state.

[0027] (5) The method provided in this application is not a simple threshold alarm, but a physical constraint-based empirical model. The method provided in this application does not only alarm when the wall-mounted quality or RMS index exceeds the threshold, but writes physical variables into empirical formulas and ensures that the parameter direction conforms to physical laws through non-negative exponential constraints. This method has both engineering interpretability and can be continuously corrected and calibrated through experimental data.

[0028] (6) The method provided in this application embodiment can be coupled with existing acoustic emission particle size prediction and instability early warning systems. The sampling time determined by the method provided in this application embodiment can be used as the input window for subsequent acoustic emission signal analysis. Wavelet packet energy analysis, RMS risk feature extraction, particle size distribution prediction, or defluidization risk early warning can then be performed, thereby forming a closed-loop control process of spraying-drying-sampling-analysis-feedback. Attached Figure Description

[0029] Figure 1 Comparison of particle adhesion to the wall under different spraying strategies provided in the embodiments of this application (Intermittent spray: intermittent spray, Continuous spray: continuous spray); Figure 2 This is a schematic diagram of the main structure of the spray fluidized bed and the acoustic signal acquisition system provided in the embodiments of this application; wherein Figure 2 (a) is a spray fluidized bed experimental device coupled with an acoustic emission acquisition system (the fluidized bed adopts a conical design with a narrow cross-section region at the bottom and a wide cross-section region at the top). Figure 2 (b) Shows the sensor installation status; Figure 3 A schematic diagram of the adaptive control method for intermittent spray-drying in a spray fluidized bed provided in this application embodiment is provided for this application embodiment; Figure 4 Fitting graph of predicted wall-mounted quality provided in the embodiments of this application; Figure 5 Relative error diagrams for various operating conditions provided in the embodiments of this application; Figure 6 This is a schematic diagram illustrating the determination of recommended sampling time for acoustic emission provided in an embodiment of this application; Figure 7 Time diagrams for predicting continuous spraying under different operating conditions provided in embodiments of this application; Figure 8 The real-time RMS values ​​of acoustic signal timing results collected under different working conditions during the spray granulation process are provided in the embodiments of this application. Figure 9 The influence of gradually increasing wall-mounted layer mass on the frequency distribution of acoustic signals under different continuous spraying conditions provided in the embodiments of this application; Figure 10 This is a schematic diagram of the structure of the spray fluidized bed intermittent spray-drying adaptive control device provided in the embodiments of this application; Figure 11 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0030] The technical solutions of 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. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0031] In the description of this application, it should also be noted that, unless otherwise expressly specified and limited, the terms "set up," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this technology based on the specific circumstances.

[0032] In the description of this application, spatial relation terms such as "below," "under," "below," "below," "above," "over," etc., are used herein to describe the relationship between one element or feature shown in the figures and other elements or features. It should be understood that, in addition to the orientation shown in the figures, spatial relation terms also include different orientations of the device in use and operation. For example, if the device in the figures is flipped, an element or feature described as "below" or "under" or "below" of other elements or features will be oriented "above" other elements or features. Therefore, the exemplary terms "below" and "under" can include both upper and lower orientations. Furthermore, the device may also include other orientations (e.g., rotated 90 degrees or other orientations), and the spatial descriptive terms used herein are interpreted accordingly.

[0033] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0034] The existing spray fluidized bed granulation process has the following technical problems: (1) Existing spray fluidized beds usually lack formulas for calculating the risk of wall adhesion, and cannot directly determine whether the single spray time is too long based on the binder viscosity, spray rate, bed size and fluidization intensity.

[0035] (2) The amount of wall adhesion formed per unit time varies under different viscosities, spray rates and bed dimensions. If a uniform spraying time is used, production efficiency will be reduced under low-risk conditions, while serious wall adhesion may be formed under high-risk conditions.

[0036] like Figure 1 The image shows a comparison of particle adhesion to the spray wall under different spraying strategies. The left image represents intermittent spraying, and the right image represents continuous spraying. The comparison clearly shows that continuous spraying results in significantly more severe particle adhesion to the spray wall than intermittent spraying.

[0037] (3) Acoustic emission acquisition is easily affected by spray droplets and wet walls.

[0038] (4) The problem of fixed drying time and lack of feedback at the sampling time.

[0039] (5) Solve the problem of lack of physical constraints in traditional threshold alarm methods.

[0040] In view of this, this application provides an adaptive control method and apparatus for intermittent spray-drying in a spray fluidized bed.

[0041] First, we will introduce the physical device on which the method provided in the embodiments of this application is applied.

[0042] See Figure 2 This is a schematic diagram of the main structure of the spray fluidized bed and the acoustic signal acquisition system provided in the embodiments of this application. Figure 2 (a) is a spray fluidized bed experimental device coupled with an acoustic emission acquisition system (the fluidized bed adopts a conical design with a narrow cross-section region at the bottom and a wide cross-section region at the top). Figure 2 (b) shows the sensor installation, demonstrating that the method is very simple to implement and has low installation requirements. It is particularly suitable for reactors with high sealing requirements, such as those operating under high pressure and high temperature.

[0043] Specifically, the top-spray granulation fluidized bed described in this application employs alternating "spraying section" and "pure drying section" operations during the granulation process, replacing the traditional continuous spraying. In the spraying section, a coagulant is sprayed to promote particle growth, followed immediately by the drying section. Hot air is used to rapidly evaporate the liquid on the particle surface and reactor wall, and fluidized collisions wash away the material layer adhering to the wall. The core of this design lies in the fact that the acoustic emission monitoring system triggers signal acquisition and data analysis only within the drying section. Because the direct interference from droplet spray is eliminated during the drying period, and the wall surface returns to a dry state, the collision energy absorption and signal shielding problems caused by excessively thick liquid films are eliminated. This results in more concentrated, more regular acoustic signal energy generated by particle collisions, and significantly reduced data fluctuations. This strategy actively captures data from the quasi-dry state period with the highest signal-to-noise ratio, providing minimally fluctuating and highly repeatable raw input for subsequent algorithms from a physical source, thereby significantly improving the accuracy of online particle size distribution calculations and the system's steady-state monitoring capabilities. This solves the problem of signal distortion and misjudgment of detection results caused by particle adhesion to the wall when applying acoustic emission technology to spray fluidized bed processes.

[0044] See Figure 3 The spray fluidized bed intermittent spray-drying adaptive control method provided in this application includes: S301. Obtain experimental data on the intermittent spray-drying process of the spray fluidized bed under different operating conditions; the experimental data includes process parameter data and wall adhesion mass; S302. Based on the experimental data, construct a predictive model that reflects the relationship between the wall hanging quality and process parameter data; S303. Construct an empirical model that reflects the relationship between drying waiting time and RMS recovery characteristics; S304. Determine the critical wall-hanging quality of the section to be controlled; S305. Based on the critical wall adhesion mass and prediction model, calculate the allowable spraying time for a single spraying cycle. S306. Based on the single allowable spray time and the empirical model, calculate the predicted drying waiting time and determine the acoustic signal sampling time.

[0045] In one possible implementation, in S301, the process parameters include binder viscosity, spray rate, bed dimensions, bed height, and fluidizing gas velocity.

[0046] In one possible implementation, in S302, the formula for the prediction model is as follows: (1) In the formula, This is a dimensionless predicted value for the wall-mounted mass. The wall-mounted quality coefficient; The Ohnesorge number is the number of the binder droplet, used to characterize the relative interaction between binder viscosity, surface tension, and droplet characteristic size; This is a dimensionless liquid load parameter used to characterize the amount of liquid sprayed per unit bed mass; This is a fluidization intensity parameter, used to characterize the ratio between the apparent fluidization velocity and the minimum fluidization velocity; , , They are respectively , , The reference values ​​are a, b, and c, which are physical constraint indices.

[0047] Furthermore, Oh L Π L and Π U The calculation formula is as follows:

[0048] In the formula, This refers to the viscosity of the adhesive. The density of the liquid; d represents the surface tension of the liquid. 32 The average diameter of the droplets (Sauter).

[0049] In the formula, L t is the mass flow rate of the sprayed liquid. s For continuous spraying time or single spraying time; M b Equivalent quality of the bed layer;

[0050] In the formula, U g U represents the apparent fluidization velocity. mf This represents the minimum fluidization rate.

[0051] In one possible implementation, in S303, the formula of the empirical model is as follows: (2) In the formula, The drying time is dimensionless. This is the drying waiting time. This corresponds to the spray duration of the cycle. This is the drying time coefficient. Let represent the decrease in spray impact during the i-th cycle, where α, η, and ζ are non-negative exponents, and ε is a minimum value to prevent the denominator from being zero. This represents the drying recovery coefficient for the (i-1)th cycle. This indicates the risk of a decrease in the later stages of drying in the (i-1)th cycle.

[0052] It should be noted that a cycle refers to a process of spraying followed by drying. In the i-th cycle, the spraying segment is denoted as the i-th spray, and the drying segment is denoted as the i-th drying.

[0053] In one possible implementation, in S305, the formula for calculating the single permissible spray time is as follows: (3) In the formula, This refers to the permissible spray time per application. For a single permissible liquid load, , For the dimensionless wall-mounted mass allowed in a single instance, , The critical wall-mounting mass; For bed equivalent quality, This represents the mass flow rate of the sprayed liquid.

[0054] In one possible implementation, in S306, the formula for calculating the predicted drying time is: ; Predict the drying time for the i-th cycle.

[0055] In one possible implementation, S306 includes: S306a. The single allowable spray time is taken as the spray duration of the section to be controlled. S306b. Calculate and predict the drying waiting time using the empirical model. The calculation formula is: ; Predict the drying waiting time for the i-th cycle. The duration of the spray in the i-th cycle; S306c. Taking the end of spraying as the starting point, and adding the predicted drying waiting time and safety delay time, the resulting time is determined as the sound signal sampling time.

[0056] Furthermore, the formula for calculating the sampling time of the acoustic signal is as follows: (4) In the formula, Let i be the sampling time of the i-th cyclic acoustic signal. The time at which the i-th spray cycle ends. Predict the drying time for the i-th cycle. The time was delayed for safety reasons.

[0057] The following detailed description is based on specific embodiments.

[0058] An adaptive control method for intermittent spray-drying in a spray fluidized bed includes the following steps: S1. Obtain experimental data on the intermittent spray-drying process of a spray fluidized bed under different operating conditions. To establish a formula for calculating spray time, continuous spraying experiments were first conducted under several typical operating conditions, or existing continuous spraying experiment results were retrieved. The wall adhesion mass was recorded under different binder viscosities, spray rates, bed dimensions, bed heights, and fluidizing gas velocities.

[0059] The specific procedure involves setting different spray times for each experiment. After the experiment, the waste material in the fluidized bed is cleaned up, taking care not to damage the wall-mounted layer. After cleaning, the adhering wall-mounted layer particles are scraped off with a brush and collected with a handheld vacuum cleaner. Each batch is weighed three times, with data fluctuations within ±0.3g.

[0060] S2. Based on the experimental data, construct a predictive model reflecting the relationship between wall adhesion quality and process parameter data. The steps for building a prediction model are as follows: (1) Express the wall-mounted mass in a dimensionless form as the equivalent mass of the bed layer: (5) In the formula, For dimensionless wall-mounted mass, For the quality of wall hanging, The equivalent quality of the bed layer.

[0061] (6) In the formula, The bed packing density, The average diameter or characteristic diameter of the fluidized bed. This refers to the static particle bed height.

[0062] It should be noted that the equivalent quality of the bed layer It is a quality standard used to characterize the magnitude of solid materials in the bed, mainly used for dimensionless processing of wall-mounted mass and spray volume. When the actual load volume is known, This can be taken as the actual mass of the solid particles inside the bed; when the actual load amount is not obtained, Based on bed packing density characteristic diameter of bed body or bed layer and feature height Estimation, i.e., formula (6). It does not represent the apparent weight of the bed after airflow lifting, but rather the size of the bed solids involved in the spray granulation process.

[0063] (2) Calculation of relevant dimensionless parameters (7) In the formula, Π L It is a dimensionless liquid load parameter used to characterize the amount of liquid injected per unit bed of solid material during spraying. This refers to the mass flow rate of the sprayed liquid. This refers to either continuous spraying time or single spraying time. The equivalent quality of the bed layer.

[0064] (8) In the formula, Π U U is a fluidization intensity parameter used to characterize the apparent fluidization velocity relative to the minimum fluidization velocity; g U represents the apparent fluidization velocity. mf This represents the minimum fluidization rate.

[0065] (9) In the formula, Oh L The Ohnesorge number is the number of the binder droplet, used to characterize the relative interaction between binder viscosity, surface tension, and droplet characteristic size; This refers to the viscosity of the adhesive. The density of the liquid; d represents the surface tension of the liquid. 32 The average diameter of the droplets is the Sauter mean.

[0066] (3) Based on the above dimensionless parameters, establish a formula for predicting the quality of wall adhesion: (10) In the formula, For dimensionless wall-mounted prediction quality, The wall-mounted quality coefficient is... , , For reference values ​​corresponding to dimensionless variables, a, b, and c are physical constraint indices. Where a ≥ 0 indicates that the tendency to adhere to the wall does not decrease with increasing viscosity-related parameters. b ≥ 0 indicates that the wall adhesion mass does not decrease with increasing liquid load. c ≥ 0, and acts as a negative exponent. This indicates that the wall adhesion quality decreases as the fluidization intensity increases.

[0067] It should be noted that, , , , a, b, and c were all obtained by log-space least squares fitting with non-negative exponential constraints.

[0068] Specifically, the methods of obtaining it include: (1) Collect continuous spray wall adhesion mass data under different binder viscosities, spray rates and fluidization conditions, and convert the wall adhesion mass into dimensionless wall adhesion mass; (2) Calculate the corresponding dimensionless liquid load parameters, fluidization intensity parameters and dimensionless viscosity parameters based on the equivalent mass of the bed, the mass flow rate of the sprayed liquid, the spraying time, the fluidizing gas velocity and the minimum fluidization velocity; (3) Take the logarithm of the wall hanging quality prediction formula and construct a linearized logarithmic space fitting relationship; (4) Under the physical constraints that a, b, and c are all not less than 0, the wall hanging quality coefficient, reference parameters, and physical constraint index are obtained by using the least squares method; (5) The fitting results are evaluated based on the fitting residuals, coefficient of determination, root mean square error and average relative error, and the obtained parameters are used to back-calculate the allowable spray time for subsequent single spraying.

[0069] In this embodiment, log-space least squares fitting with non-negative exponential constraints is used to obtain the parameters shown in Table 1.

[0070] Table 1. Parameter Fitting Values

[0071] Figure 4 The fitting results of equation (10) are shown: R²=0.9578, RMSE=1.9511 g, MAE=1.5109 g, MAPE=6.7907%.

[0072] Figure 5 The relative error diagrams for each operating condition, obtained based on equation (10), are shown. It can be seen from the diagrams that... Figure 5 The relative error diagrams for each operating condition obtained based on Equation (10) are shown. As can be seen from the diagrams, the relative errors between the predicted and experimental results under each operating condition are generally at a low level, ranging from approximately -12.5% ​​to 11.8%. Specifically, the relative errors for the 38.1 mPa·s, 1 g / min and 413.7 mPa·s, 1 g / min conditions are close to 0, indicating that the model has a good fit for both low-viscosity and high-viscosity baseline conditions. The error for the 293.6 mPa·s, 1 g / min condition is positive, while the error for the 413.7 mPa·s, 0.8 g / min condition is negative, indicating that the model still has some deviations under some intermediate operating conditions, but the overall error does not show a systematic shift in a single direction. The above results show that Equation (10) can well reflect the changing trend of the wall-attached mass under different viscosities and spray rates, and can be used for subsequent calculations of the allowable spray time per spray.

[0073] S3. Construct an empirical model reflecting the relationship between drying waiting time and RMS recovery characteristics. (1) Parameter definition In each spray-dry cycle, an acoustic emission sensor acquires the elastic wave signal generated by particle collisions and calculates the RMS characteristic value.

[0074] The RMS signal typically decreases or fluctuates in the spraying section, reflecting the effects of droplet injection, particle wetting, impact damping, and wall adhesion on the acoustic signal. Upon entering the drying section, the RMS signal gradually recovers as liquid on the particle surface evaporates and wall residue weakens.

[0075] (a) Define the spray impact drop amplitude : (11) In the formula, Let RMS mean be the initial stage value of the i-th spray segment. This represents the RMS mean of the low-value area in the spray section. The larger the value, the more significant the disturbance of the spray on the bed's collision sound signal.

[0076] Specifically, the initial stage of the i-th spray segment is the time period corresponding to the first 10% of the spray duration starting from the start of the spray.

[0077] Specifically, the low-value region of the spray segment is a continuous time window centered on the moment corresponding to the minimum RMS value of the spray segment, with a duration of 10% of the spray duration. If this window exceeds the boundary of the spray segment, it is truncated according to the actual boundary of the spray segment.

[0078] (b) Define the drying recovery coefficient This coefficient is used to characterize the degree to which the RMS signal in the drying section recovers from a low spray value to a stable state. The larger the value, the more complete the drying and recovery process.

[0079] Specifically,

[0080] In the formula, The RMS average value of the drying stable period before the start of the i-th spray is used as the baseline value for the stable state before spraying. is the RMS average value in the low-value region of the i-th spray segment, used to characterize the state when the signal drops to a low level after the spray disturbance; The RMS average value within the preset time window at the end of the i-th drying stage (i.e., the last 10% of the drying stage) is used to characterize the recovery level of the signal before the end of drying. To prevent the introduction of extremely small positive numbers due to a denominator of zero, it is recommended to use 10. -8 .

[0081] (c) Define the risk of decline in the later stages of drying. This parameter reflects whether the RMS still shows a downward trend at the end of the drying stage. It is obtained by calculating the slope of the RMS change over time within the window at the end of the drying stage and combining it with the proportion of continuously decreasing points. If the signal continues to decline in the later stage of the drying stage, it indicates that the bed or wall condition may not yet be stable, and the waiting time needs to be extended.

[0082] Specifically,

[0083] in

[0084] In the formula, The slope of the fitted RMS change over time within the window at the end of the i-th drying stage (the last 10%) is the slope of the fitted RMS change over time. The duration of this final window; This represents the average RMS value within the final time window. This represents the total number of RMS data points within the final window. This represents the number of times the RMS value at a later time step is lower than the RMS value at the previous time step among adjacent data points. This represents the proportion of points that have fallen; The normalized decrease is calculated based on the final slope; ε is a minimal positive number introduced to prevent the denominator from being zero, and is recommended to be 10. -8 .

[0085] For example, if data point N=6, then there are 6 points, which is the time series. The time is 1-6 seconds, the time interval is 1 second, and the corresponding RMS data are 10, 9, 8, 8.5, 8, 7.5 respectively.

[0086] Starting from the second point (10), the value is compared with the previous point, and the comparison results are as follows: 9 is lower than 10: a decrease; 8 is lower than 9: a decrease; 8.5 is higher than 8: No decrease; 8 is lower than 8.5: a decrease; 7.5 is lower than 8: a decrease.

[0087] Therefore, in this example, Ntail,i-1=5. Ndown, i=4, the number of descents is 4.

[0088] Therefore, the calculated 4 / 5 = 0.8, which means the decrease rate is 80%. This mainly indicates whether the RMS in the final stage of drying goes up or down, depending on the stability of the bed.

[0089] In the calculation, firstly, a linear fit is performed on RMS and time within the window at the end of the drying period to obtain the slope; secondly, the number of decreases in adjacent RMS data points within this window is counted, and the proportion of decrease points is calculated; then, the RMS decrease corresponding to the slope is divided by the average RMS of this window to obtain the normalized decrease. Finally, by multiplying the two results, we obtain the risk of a decrease in the later stages of drying. Note here, when When ≥0, Setting it to 0 can be interpreted as an artificial cutoff, indicating that there is no risk of decline at the end of the drying process; The larger the value, the more pronounced the downward trend at the end of the drying process.

[0090] (2) Construct an empirical model based on the above parameters To determine the sampling time of the acoustic signal after spraying, this application establishes an empirical formula for the drying waiting time.

[0091] Define dimensionless drying waiting time : (12) In the formula, This is the drying waiting time. This represents the spray duration for the corresponding cycle.

[0092] It should be noted that the drying waiting time This refers to the time required from the end of the spray until the acoustic signal recovers to a usable level or reaches the preset drying recovery criterion. It corresponds to the spray duration of the cycle. This refers to the duration of spraying within the same spray cycle.

[0093] In the offline experience summary mode, the embodiments of this application use the following formula to describe the relationship between drying waiting time and RMS recovery characteristics: (13) In the formula, This is the drying waiting time. This corresponds to the spray duration of the cycle. This is the drying time coefficient. The values ​​α and η represent the decrease in spray impact during the i-th cycle (this cycle). ε is a non-negative exponent, and ε is a minimal quantity to prevent the denominator from being zero. This represents the drying recovery coefficient for the (i-1)th cycle. This indicates the risk of a decrease in the later stages of drying in the (i-1)th cycle. The more pronounced the decrease in acoustic signal caused by spraying, the longer the drying time. This indicates that the lower the recovery coefficient, the longer the drying time. This indicates that the higher the risk of decline in the later stages of the previous drying cycle, the longer the waiting time for this drying cycle will be.

[0094] To facilitate the direct calculation of the predicted drying time, equation (13) can be equivalently transformed into In the formula, Predict the drying waiting time for the i-th cycle. The duration of the spray in the i-th cycle.

[0095] In this embodiment, equation (13) is calibrated offline using data from adjacent spray-dry cycles. For the i-th cycle, where i ≥ 2, the i-th cycle is used as the basis for calibration. 1 completed loop , and As input to the model, the actual dimensionless drying waiting time of the i-th cycle. As a benchmark. Among them, , Let be the actual duration of the drying segment in the i-th cycle, and = ; Let i be the start time of the drying segment in the i-th cycle. The end time of the i-th drying segment. The actual spray duration for the i-th cycle.

[0096] After constructing the adjacent cycle samples based on the RMS dataset, offline empirical fitting was performed on equation (13), and the file hold-out method was used for verification. The number of verification files was 200, the average RMSE was 40.5249 s, the average MAE was 31.909 s, and the average MAPE was 10.2637%. Since the first cycle does not have the RMS features of the previous cycle, it does not participate in the construction of adjacent cycle samples, and its drying waiting time is determined using a pre-calibrated initial value.

[0097] It should be noted that, and All of these characteristics were calculated from data from the (i-1)th completed spray-dry cycle. Therefore, at the end of the i-th spray cycle, all of the above characteristics are known quantities and can be used to calculate the predicted drying waiting time for the i-th cycle. To avoid using complete dry segment data that has not yet been obtained in the current cycle, The larger the value, the more significant the disturbance to the acoustic emission signal from the previous spray cycle, and the longer the predicted drying waiting time for the current cycle. The smaller the value, the worse the drying recovery in the previous cycle, and the longer the predicted drying waiting time in the current cycle. The larger the value, the greater the risk of decline in the later stages of the previous drying cycle, and the longer the predicted drying waiting time for the current cycle.

[0098] For the first spray-dry cycle, since the RMS characteristics of the previous cycle are not present, the pre-calibrated initial drying waiting time under the same or similar operating conditions is used. As the predicted drying waiting time for the first cycle, i.e. = Equation (13) transforms the RMS disturbance and recovery state of the previous cycle into the drying waiting time of the current cycle, thereby enabling adaptive updates of drying time between different cycles.

[0099] Figure 6 This is a schematic diagram of determining the recommended sampling time based on equation (13). The solid line represents the end time of spraying in each cycle, and the dashed line represents the recommended sampling time. The time interval between the two consists of the predicted drying time and a 1-second safety delay. This figure illustrates that the acoustic emission signal is not collected immediately after the spraying ends, but after the bed has dried and recovered to a certain extent, thereby reducing the interference of droplet spray and wet wall residue on the acoustic signal.

[0100] It can be understood that the method for calculating the RMS eigenvalue includes: 1. Signal Acquisition Converting continuous analog sound waves into discrete digital signals that can be processed by a computer.

[0101] Generate the original time series x[n], where n takes values ​​of 0, 1, 2, ..., N-1, and N represents the total number of sampling points.

[0102] 2. Signal preprocessing Eliminate clutter interference in the original signal and improve the accuracy of RMS feature calculation.

[0103] Preprocessing methods include bandpass digital filtering.

[0104] 3. Windowing and framing (used to generate temporal dynamic RMS) Sound waves are non-stationary signals, but can be approximated as stationary signals within a short time interval of 20~40ms. Therefore, the signal is split into independent short time frames.

[0105] 3.1 Frame length L Commonly used sampling point numbers are 512, 1024, and 2048; for example, at a sampling rate of 44.1kHz, 1024 points correspond to a duration of approximately 23.2ms.

[0106] 3.2 Frame Shift The overlap length between adjacent frames is typically set to 50% or 25% of the frame length to smooth the output feature curve.

[0107] 3.3 Windowed Operations To suppress spectral leakage, each frame of signal is multiplied by a window function, commonly the Hamming window or Hanning window; Calculation formula: xwin[m] = xframe[m]·w[m], where m takes values ​​of 0, 1, ..., L 1.

[0108] 4. Calculation of RMS feature value of a single frame RMS is used to characterize the effective energy and average amplitude of a frame of signal.

[0109] 4.1 Basic Formula for RMS in Linear Domain (14) In equation (14), RMS is the root mean square value in the linear domain of the current signal frame, which is used to characterize the effective amplitude or energy level of the acoustic signal in that frame; L is the number of sampling points contained in a single frame; This represents the m-th sampled value of the current signal frame after processing by the window function; m is the index of the sampled point within the frame, taking values ​​of 0, 1, 2, ..., L. 1.

[0110] The calculated numerical units are consistent with the original sampling units (voltage V, sound pressure Pa, etc.), but the numerical range is too small.

[0111] 4.2 Logarithmic domain decibel RMS (most commonly used in engineering) Aligned with human auditory perception, the conversion formula is as follows: (15) In equation (15), RMS is the root mean square value expressed in decibels; RMS is the root mean square value in the linear domain calculated according to equation (14); REF is the reference amplitude, which can be taken as the standard reference sound pressure in air (20 μPa) when calibrating using physical sound pressure, and when using an amplitude range of For normalized digital signals of 1 to 1, 1.0 can be used; ε To prevent the introduction of extremely small positive numbers due to the meaninglessness of logarithmic operations when RMS is zero, this embodiment can take 10. -8 .

[0112] 4.2.1 REF reference value: When calibrating physical sound pressure, take the standard air reference sound pressure of 20μPa; take 1.0 for pure digital normalized signal (amplitude -1~1).

[0113] 4.2.2 ε is the minimum offset constant, typically taken as 10. -8 This prevents errors caused by inputting zero values, which would result in an error message indicating that the logarithm is meaningless.

[0114] 4.3 Explanation of Window Function Energy Compensation Non-rectangular windows such as the Hamming window and Hanning window will attenuate the total signal energy; for scenarios with strict requirements for energy accuracy, compensation coefficients need to be superimposed after the calculation, and the value of the coefficients is equal to the reciprocal of the sum of the squares of all coefficients of the window function.

[0115] 5. Sliding frame batch calculation of timing RMS sequence For the entire complete signal, repeat the framing, windowing, and RMS operations with a fixed frame shift: 5.1 Frame Reading Rules Frame 1: Sampling points [0] to [L] 1] Frame 2: Sampling point [shift] to [shift+L] 1] Subsequent frames slide offset sequentially 5.2 Output Sequence Specifications The final output is a one-dimensional time-series feature sequence RMS[k], with a total sequence length approximately equal to... Where N is the total number of sampling points, L is the frame length of a single frame, and shift is the frame shift. This sequence is used to fully reconstruct the fluctuation trend of sound wave energy over time.

[0116] S4. Determine the critical wall-hanging quality of the section to be controlled.

[0117] This step determines the critical wall-hanging quality based on the actual conditions of the section to be controlled.

[0118] In one possible implementation, the critical wall-attachment mass can be obtained through experimental calibration or empirical data.

[0119] Understandably, the critical wall-mount mass serves to set an allowable upper limit for the amount of wall-mounted material: when the wall-mounted mass is below this upper limit, the influence of the wall-mounted deposit on the acoustic signal is still within an acceptable range; when the wall-mounted mass exceeds this upper limit, the absorption and damping effect of the deposit on the acoustic wave is significantly enhanced, which may lead to unacceptable deviations in RMS, frequency band energy distribution, or subsequent particle size prediction results.

[0120] For example, the method for determining the critical wall adhesion mass through experimental calibration is as follows: (1) With the bed structure, sensor installation position and fluidized bed conditions remaining unchanged, different continuous spraying times were set to form different deposition layers on the wall from thin to thick. After each set of experiments, the corresponding wall adhesion mass was weighed and the acoustic emission signal under that condition was collected. Subsequently, the RMS, characteristic frequency band energy ratio or particle size prediction results corresponding to each wall adhesion mass were compared with the benchmark results under the clean wall or low wall adhesion conditions.

[0121] (2) When the wall-mounted mass increases to a certain value, if the deviation of the acoustic signal from the reference signal exceeds the preset allowable range for the first time, or the particle size prediction error obtained from the acoustic signal exceeds the allowable error for the first time, Then the value is determined as Critical wall adhesion quality If the critical state is located between two adjacent experimental points, it can be determined by interpolation based on the relationship between the wall mass and the acoustic signal deviation.

[0122] The critical wall-attachment mass obtained from this is used to subsequently back-calculate the permissible spray time per spray, which is essentially "the maximum wall-attachment mass allowed when the acoustic signal can still be reliably used".

[0123] For example, in a certain control section of this application embodiment, the critical wall-mounted mass is determined to be 10 g.

[0124] In the experiment of this embodiment, the continuous spraying time was changed to gradually increase the mass of the wall deposit layer, and the RMS changes and frequency band energy distribution under different wall deposit masses were compared simultaneously. When the wall deposit mass was less than about 10 g, although the acoustic signal would fluctuate to some extent, the overall amplitude and main frequency band distribution were still relatively close to the low wall deposit state, and the subsequent acoustic analysis results were still within an acceptable range. When the wall deposit mass increased to more than about 10 g, the absorption, attenuation, and frequency filtering effect of the deposit layer on particle collision sound waves began to be significantly enhanced, and the RMS and characteristic frequency band energy distribution showed a significant shift. Continuing to use the acoustic signal under this state may affect the particle size prediction, flow pattern identification, or instability early warning results.

[0125] Therefore, in this embodiment, approximately 10 g is used as the critical wall adhesion mass under the current device and sensor arrangement to distinguish between two states where the acoustic signal is less affected and more significantly affected. It should be noted that 10 g is the calibration result of this embodiment and is not a universal constant applicable to all fluidized beds. For different bed sizes, wall materials, sensor installation locations, or material systems, the corresponding critical wall adhesion mass can be re-determined according to the aforementioned wall adhesion mass-acoustic signal response calibration method.

[0126] S5. Based on the critical wall adhesion mass and prediction model, calculate the allowable spray time for a single spraying cycle. The critical wall-attachment mass is set as: (16) And a conservative correction coefficient is introduced. This is to account for the possibility that the actual wall-mounted mass may be higher than the weighing result. In this embodiment, we take... .

[0127] (17) In the formula, The predicted mass of the wall-mounted material is equivalent to the predicted weighing result of the wall-mounted material.

[0128] Dimensionless transformation of both sides of equation (17) is: ; The critical dimensionless wall-mounted mass is: (18) In equation (10) Taken as the critical dimensionless wall-mounted mass and combined , With the reference parameters, the single allowable liquid load can be solved by inversely from the formula for predicting the wall-mounted mass: (19) Furthermore, according to equation (7) ,Will Converted to recommended single spray time: (20) It should be noted that in equation (5) This is used to convert the wall-mounted mass into a dimensionless form (i.e., the definition), and as the dependent variable (relationship) of the prediction model in equation (10). When calculating the spraying time, first... The critical dimensionless wall-mounted mass was calculated. As the dependent variable in equation (10), we obtain Finally, according to The definition was converted to obtain .therefore, It is an intermediate normalized variable that connects the wall-mounted quality prediction model with the calculation of the single allowable spray time.

[0129] Equation (20) is based on the known equivalent mass of the bed. Mass flow rate of sprayed liquid and allowable liquid load Then, the maximum duration of a single spray under the current operating conditions can be calculated. The larger the value, the higher the liquid volume the bed can withstand and the longer the allowable spraying time per spray. The larger the value, the more liquid is injected per unit time, and the shorter the allowable spray time per burst under the same allowable liquid load. This formula can calculate the single spray time under different operating conditions based on binder viscosity, spray rate, bed size, and fluidization intensity. For high viscosity and high spray rate conditions, the formula will give a shorter spray time, thereby reducing wall adhesion; for low viscosity or low spray rate conditions, the formula allows for a longer spray time to balance production efficiency.

[0130] Figure 7 The predicted continuous spraying time is shown under different operating conditions.

[0131] S6. Based on the single allowable spray time and the empirical model, calculate the predicted drying waiting time and determine the acoustic signal sampling time.

[0132] In one possible implementation, S6 includes: S6a. The single allowable spray time is taken as the spray duration of the section to be controlled; S6b. Calculate and predict the drying waiting time using the empirical model. The calculation formula is: ; Predict the drying waiting time for the i-th cycle. The duration of the spray in the i-th cycle; S6c. Taking the end of spraying as the starting point, and adding the predicted drying waiting time and safety delay time, the resulting time is determined as the sound signal sampling time.

[0133] Specifically, Allowable spray time per spray Spray duration as the control section After each spraying, the predicted drying waiting time is calculated according to equation (13). In actual control, the start time of spraying is recorded as... The spraying ends at the time when Subsequently, calculations were performed based on the spray section and drying recovery characteristics. And further introduce a safety delay time ,For example Then the sampling time of the recommended sound signal in the i-th cycle is: (twenty one) In the formula, Let i be the sampling time of the i-th cyclic acoustic signal. The time at which the i-th spray cycle ends. Predict the drying waiting time for the i-th cycle. The time was delayed for safety reasons.

[0134] It is understood that the drying waiting time is determined by an empirical model corresponding to the acoustic emission RMS recovery characteristics of the drying section, and a certain time is delayed after the drying waiting time ends as the recommended sampling time. This drying waiting time is used to allow the residual liquid in the bed and wall to fully recover, avoiding the collection of acoustic emission signals affected by droplet impact and wet walls immediately after the spray ends.

[0135] In practical applications, the system can set a short sampling window, such as 1 to 3 seconds, around this time to calculate RMS, band energy, wavelet packet energy distribution, or other acoustic features. Since this sampling time avoids the direct disturbance of the spray and the initial wet recovery stage, the obtained acoustic signal is more suitable for particle size prediction, flow regime identification, or instability early warning.

[0136] The intermittent spraying provided in this application embodiment, taking the i-th cycle as an example, has the following workflow: spraying start → spraying duration. → Spraying ends → Wait for drying → Delay → Acoustic signal sampling → Enter the next spray cycle.

[0137] See Figure 8 Under the same binder viscosity, the real-time RMS changes corresponding to different spraying rates showed significant differences. As the spraying rate increased from 0.6 g / min to 1.0 g / min, the disturbance to the acoustic emission signal during the spraying phase intensified, and the fluctuation amplitude of the RMS curve increased. After spraying and entering the drying phase, the signal required a certain amount of time to recover to a relatively stable state. This result indicates that the spraying rate affects the bed wettability and wall adhesion, thereby altering the intensity and stability of the particle collision acoustic signal. Therefore, acquiring the acoustic signal immediately after spraying is easily affected by spray disturbances and the wet recovery process, while determining the sampling time after drying recovery is more conducive to obtaining stable and repeatable acoustic emission signals.

[0138] See Figure 9 Different continuous spraying times correspond to different wall deposition layer qualities and thicknesses, and the formation of the deposition layer alters the frequency energy distribution of the acoustic signal. As the spraying time increases, the wall deposition layer gradually thickens. The sound waves generated by particles colliding with the wall are absorbed, damped, and filtered by the deposition layer before reaching the sensor, causing a decrease in the energy proportion of some frequency bands and an increase in the energy proportion of others, resulting in a shift in the frequency distribution. This indirectly demonstrates that the wall deposition layer not only changes the amplitude of the acoustic signal but also alters its frequency domain characteristics. Directly using acoustic signals affected by the deposition layer for particle size prediction, flow pattern identification, or instability warning may lead to judgments that deviate from the actual bed state. Therefore, this application improves the reliability of subsequent acoustic monitoring results by determining intermittent spray-drying operations and recommended sampling times to avoid direct spray disturbance and the initial wet recovery stage.

[0139] The following describes the spray fluidized bed intermittent spray-drying adaptive control device provided in this application. The device described below can be referred to in correspondence with the method described above.

[0140] Figure 10This is a schematic diagram of the structure of the spray fluidized bed intermittent spray-drying adaptive control device provided in the embodiments of this application, as shown below. Figure 10 As shown, the device includes: an acquisition module 101, a first construction module 102, a second construction module 103, a determination module 104, a first calculation module 105, and a second calculation module 106; The acquisition module 101 is used to acquire experimental data of the intermittent spray-drying process of the spray fluidized bed under different operating conditions; the experimental data includes process parameter data and wall adhesion mass; The first construction module 102 is used to construct a predictive model based on the experimental data, reflecting the relationship between the wall adhesion quality and process parameter data. The second building module 103 is used to build an empirical model that reflects the relationship between drying waiting time and RMS recovery characteristics; Module 104 is used to determine the critical wall-hanging quality of the section to be controlled; The first calculation module 105 is used to calculate the allowable spraying time for a single spraying based on the critical wall-mounting mass and the prediction model. The second calculation module 106 is used to calculate the predicted drying waiting time based on the single allowable spray time and the empirical model, and to determine the sampling time of the acoustic signal.

[0141] Figure 11 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 11 As shown, the electronic device may include a processor 1110, a communications interface 1120, a memory 1130, and a communications bus 1140, wherein the processor 1110, the communications interface 1120, and the memory 1130 communicate with each other through the communications bus 1140. The processor 1110 can call logic instructions in the memory 1130 to execute an adaptive control method for intermittent spraying-drying in a spray fluidized bed. This method includes: acquiring experimental data under different operating conditions, constructing a wall adhesion quality prediction model and a drying waiting time empirical model, determining the critical wall adhesion quality and drying waiting time, calculating the single allowable spray time, and calculating the predicted drying time to determine the acoustic signal sampling time.

[0142] Furthermore, the logical instructions in the aforementioned memory 1130 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0143] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the spray fluidized bed intermittent spray-drying adaptive control method provided by the above methods.

[0144] In another aspect, this application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the spray fluidized bed intermittent spray-drying adaptive control method provided by the above methods.

[0145] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0146] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An adaptive control method for intermittent spray-drying in a spray fluidized bed, characterized in that, include: Experimental data on the intermittent spray-drying process of a spray fluidized bed under different operating conditions were obtained; the experimental data included process parameter data and wall adhesion mass. Based on the experimental data, a predictive model reflecting the relationship between wall adhesion quality and process parameter data is constructed. An empirical model reflecting the relationship between drying waiting time and RMS recovery characteristics was constructed. Determine the critical wall-hanging quality of the section to be controlled; Based on the critical wall adhesion mass and prediction model, the allowable spraying time for a single spraying cycle is calculated. Based on the single allowable spray time and the empirical model, the predicted drying waiting time is calculated, and the sampling time of the acoustic signal is determined.

2. The method according to claim 1, characterized in that, The process parameters include binder viscosity, spray rate, bed dimensions, bed height, and fluidizing gas velocity.

3. The method according to claim 1, characterized in that, The formula for the prediction model is as follows: In the formula, Y w C is the dimensionless predicted value of the wall-mounted mass. w For wall-mounted quality coefficient; Oh L The Ohnesorge number is the number of the adhesive droplet, used to characterize the relative interaction between adhesive viscosity, surface tension, and droplet characteristic size; Π L Π is a dimensionless liquid load parameter used to characterize the amount of liquid injected per unit bed mass; U The fluidization intensity parameter is used to characterize the ratio between the apparent fluidization velocity and the minimum fluidization velocity; Oh0, Π L0 Π U0 Oh L Π L Π U The reference values ​​are a, b, and c, which are physical constraint indices.

4. The method according to claim 1, characterized in that, The formula for the empirical model is as follows: In the formula, The drying time is dimensionless. This is the drying waiting time. This corresponds to the spray duration of the cycle. This is the drying time coefficient. Indicates the decrease in spray impact during the i-th cycle, α, η, ε is a non-negative exponent, and ε is a minimal quantity to prevent the denominator from being zero. This represents the drying recovery coefficient for the (i-1)th cycle. This indicates the risk of a decrease in the later stages of drying during the (i-1)th cycle.

5. The method according to claim 1, characterized in that, The formula for calculating the permissible spray time per spray is as follows: In the formula, This refers to the permissible spray time per application. For a single permissible liquid load, , For the dimensionless wall-mounted mass allowed in a single instance, , The critical wall-mounting mass; For bed equivalent quality, This represents the mass flow rate of the sprayed liquid.

6. The method according to claim 1, characterized in that, The step of calculating the predicted drying waiting time based on the single allowable spray time and combining it with the empirical model, and determining the acoustic signal sampling time, includes: The single allowable spray time is taken as the spray duration of the section to be controlled; The predicted drying waiting time is calculated using the empirical model described above; the calculation formula is as follows: ; Predict the drying waiting time for the i-th cycle. The duration of the spray in the i-th cycle; The moment obtained by adding the predicted drying waiting time and safety delay time to the end of the spraying time is determined as the sampling time of the acoustic signal.

7. A spray fluidized bed intermittent spray-drying adaptive control device, comprising: The acquisition module is used to acquire experimental data on the intermittent spray-drying process of a spray fluidized bed under different operating conditions; The experimental data includes process parameter data and wall adhesion quality; The first construction module is used to construct a predictive model based on the experimental data, reflecting the relationship between the wall adhesion quality and process parameter data. The second building module is used to construct an empirical model that reflects the relationship between drying waiting time and RMS recovery characteristics; The determination module is used to determine the critical wall adhesion quality of the section to be controlled; The first calculation module is used to calculate the allowable spraying time for a single spraying operation based on the critical wall-mounted mass and the prediction model. The second calculation module is used to calculate the predicted drying waiting time based on the single allowable spray time and the empirical model, and to determine the sampling time of the acoustic signal.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the spray fluidized bed intermittent spray-drying adaptive control method as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the spray fluidized bed intermittent spray-drying adaptive control method as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the spray fluidized bed intermittent spray-drying adaptive control method as described in any one of claims 1 to 6.

Images