Fluidized bed granulation process moisture endpoint determination method, device, equipment and medium based on near infrared spectrum combined with moving window standard deviation

Abstract

This invention discloses a method, apparatus, equipment, and medium for determining the moisture endpoint in fluidized bed granulation based on near-infrared spectroscopy combined with moving window standard deviation. Relating to the field of fluidized bed granulation technology, this method acquires near-infrared spectral data, material moisture content, and particle size data through a multi-modal sensing acquisition unit, and obtains a fused moisture value after fusion correction. Moving window standard deviation analysis is performed on the near-infrared spectral data to construct a spectral fluctuation characteristic function to determine the moisture distribution on the particle surface. Combining the fused moisture value and particle size data, an endpoint signal is generated through a dual-index collaborative determination logic to complete the moisture endpoint determination. This method can achieve accurate and real-time determination of the granulation moisture endpoint, and can also confirm the endpoint through two-dimensional characteristic spectra, adaptively adjust the feed rate, improve the stability of granulation quality, adapt to the needs of fluidized bed granulation production, and has high engineering application value.

Description

Technical Field

[0001] This invention relates to the field of fluidized bed granulation technology, and more specifically, to a method, apparatus, equipment, and medium for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window. Background Technology

[0002] Background of a Moisture Endpoint Determination Method in Fluidized Bed Granulation Based on Near-Infrared Spectroscopy and Moving Window Standard Deviation: Fluidized bed granulation technology is a core process for preparing granular materials in pharmaceutical, chemical, and other fields. The moisture endpoint determination directly determines particle size distribution, porosity, and the stability of the final product. Traditional determination methods mainly rely on offline sampling and drying detection or single-sensor monitoring, which have significant technical bottlenecks.

[0003] Offline detection suffers from significant time lag, failing to reflect dynamic process changes and easily leading to excessive dryness or wetness. While single near-infrared spectroscopy can only reflect surface moisture, it struggles to characterize the overall bed moisture content and is susceptible to interference from particle size variations. Furthermore, existing technologies often employ fixed threshold judgments, ignoring the nonlinear dynamic evolution of material states during granulation. Uneven moisture distribution between the surface and interior layers can easily lead to misjudgments of the endpoint, resulting in excessively wide particle size distributions or particles that are too hard or too soft, severely impacting product quality consistency. In addition, the lack of online assessment methods for moisture distribution uniformity makes it impossible to effectively warn of abnormal moisture migration during granulation, resulting in poor process controllability and low finished product yield. Therefore, there is an urgent need for a collaborative judgment technology that can simultaneously integrate multi-source data, accurately analyze moisture distribution, and is based on dynamic process characteristics. This technology would address the key issues of existing judgments, such as lag, simplification, and inaccuracy, ensuring the efficiency and stability of fluidized bed granulation. Summary of the Invention

[0004] In order to overcome the problems of delayed and singular determination and inability to accurately and collaboratively determine the moisture endpoint in existing technologies, this invention discloses a method, apparatus, equipment and medium for determining the moisture endpoint in fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of the moving window, which can effectively solve the above-mentioned technical problems.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] A method for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window, the method comprising:

[0007] Multi-source monitoring data during the fluidized bed granulation process is acquired through a multi-modal sensing acquisition unit. The multi-source monitoring data includes at least near-infrared spectral data, material moisture data, and particle size data.

[0008] The multi-source monitoring data are fused and corrected to obtain a fused moisture value;

[0009] A moving window standard deviation analysis was performed on the near-infrared spectral data to construct a spectral fluctuation characteristic function, and the surface moisture distribution of the particles was determined based on the spectral fluctuation characteristic function.

[0010] Based on the fused moisture value and particle size data, a moisture endpoint determination signal is generated through a dual-index collaborative determination logic.

[0011] Based on the moisture endpoint determination signal, the moisture endpoint of the fluidized bed granulation process is determined.

[0012] Furthermore, the multimodal sensing acquisition unit includes a near-infrared spectral acquisition module for acquiring near-infrared spectral data, a humidity sensing module for acquiring material moisture data, and a particle size monitoring module for acquiring particle size data.

[0013] Further, the process of fusing and correcting the multi-source monitoring data to obtain the fused moisture value includes:

[0014] Receive the surface moisture prediction value output by the near-infrared spectroscopy acquisition module and the overall moisture value output by the humidity sensing module;

[0015] A moisture fusion correction model is constructed based on the predicted surface moisture value and the overall moisture value. The moisture fusion correction model uses dynamic fusion weight coefficients to weight the predicted surface moisture value and the overall moisture value to obtain the fused moisture value.

[0016] The dynamic fusion weighting coefficient is dynamically adjusted according to different stages of the granulation process.

[0017] Furthermore, the step of performing moving window standard deviation analysis on the near-infrared spectral data to construct a spectral fluctuation characteristic function includes:

[0018] Moving window standard deviation analysis was performed on the full-band spectral information of near-infrared spectral data to obtain spectral absorbance values ​​and mean absorbance values;

[0019] Based on the spectral absorbance value and the average absorbance value within the moving window, a spectral fluctuation characteristic function is constructed;

[0020] Based on the changing trend of the spectral fluctuation characteristic function, it is determined whether the moisture on the particle surface tends to be uniformly distributed.

[0021] Furthermore, the step of generating a moisture endpoint determination signal based on the fused moisture value and particle size data through a dual-indicator collaborative determination logic includes:

[0022] Obtain the rate of change of fusion moisture value and the rate of change of particle size characteristic parameters;

[0023] A dual-index collaborative judgment function is constructed based on the rate of change of the fusion moisture value and the rate of change of the particle size characteristic parameter. The judgment function performs a weighted summation of the rate of change of the fusion moisture value and the rate of change of the particle size characteristic parameter.

[0024] When the calculation result of the determination function continuously meets the preset determination conditions, the moisture endpoint determination signal is triggered.

[0025] Furthermore, the method also includes:

[0026] A two-dimensional dynamic feature spectrum was constructed with granulation time as the horizontal axis and fusion moisture value and particle size characteristic parameters as the vertical axes, respectively.

[0027] A moisture endpoint determination region is defined in the two-dimensional dynamic feature spectrum. When the fused moisture value and particle size characteristic parameters simultaneously enter the determination region and meet the preset residence time requirement, a granulation endpoint confirmation signal is issued.

[0028] Furthermore, the method also includes: after receiving the moisture endpoint determination signal, generating a granulation process control command to adaptively adjust the feed rate during the granulation process.

[0029] Furthermore, a device for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window, the device comprising:

[0030] The data acquisition module is used to acquire multi-source monitoring data during the fluidized bed granulation process through a multi-modal sensing acquisition unit. The multi-source monitoring data includes at least near-infrared spectral data, material moisture data, and particle size data.

[0031] The fusion correction module is used to perform fusion correction processing on the multi-source monitoring data to obtain a fused moisture value;

[0032] The deviation analysis module is used to perform moving window standard deviation analysis on the near-infrared spectral data, construct a spectral fluctuation characteristic function, and determine the moisture distribution state on the particle surface based on the spectral fluctuation characteristic function.

[0033] The collaborative determination module is used to generate a moisture endpoint determination signal based on the fused moisture value and particle size data through dual-index collaborative determination logic.

[0034] The endpoint determination module is used to determine the moisture endpoint in the fluidized bed granulation process based on the moisture endpoint determination signal.

[0035] Furthermore, an electronic device, the electronic device comprising:

[0036] Memory containing executable program code;

[0037] A processor coupled to the memory;

[0038] The processor calls the executable program code stored in the memory to execute the method for determining the moisture endpoint in the fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of the moving window, as described above.

[0039] Furthermore, a computer storage medium stores computer instructions that, when invoked, execute the method described above for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window.

[0040] Compared with the prior art, the beneficial effects of the present invention are: Attached Figure Description

[0041] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other embodiments can be derived from the provided drawings without creative effort.

[0042] Figure 1 A flowchart of a method for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window, provided for an embodiment of this application;

[0043] Figure 2 A structural diagram of a fluidized bed granulation process moisture endpoint determination device based on near-infrared spectroscopy combined with moving window standard deviation provided in this application embodiment;

[0044] Figure 3 This is a schematic diagram of an electronic device structure provided in an embodiment of this application. Detailed Implementation

[0045] The accompanying drawings are for illustrative purposes only and should not be construed as limiting the scope of this patent.

[0046] To better illustrate this embodiment, some parts in the accompanying drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions;

[0047] It will be understood by those skilled in the art that certain well-known structures and their descriptions may be omitted in the accompanying drawings.

[0048] The embodiments of this application will be further described in detail below with reference to the accompanying drawings and examples.

[0049] It is understood that the specific embodiments described herein are merely illustrative of the embodiments of this application and are not intended to limit the embodiments of this application. Furthermore, it should be noted that, for ease of description, the accompanying drawings only show the parts related to the embodiments of this application, not all structures. Those skilled in the art, after reading this specification, should be able to realize that any combination of technical features can constitute an optional implementation method, provided that the technical features do not contradict each other.

[0050] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship. In the description of this application, "multiple" means two or more, and "several" means one or more.

[0051] The moisture endpoint determination method for fluidized bed granulation based on near-infrared spectroscopy combined with moving window standard deviation provided in this application can be applied to fluidized bed granulation production scenarios for various materials in pharmaceutical, chemical, and food industries. Through multi-source data fusion correction, spectral fluctuation analysis, and dual-index collaborative determination, it achieves accurate and real-time determination of the moisture endpoint, solving the technical pain points of lagging, inaccurate, and singular determination in existing technologies, and improving the controllability of the granulation process and the stability of product quality. This application does not limit the specific application scenarios.

[0052] The method for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window provided in this application embodiment can be executed by an electronic device. The electronic device refers to any electronic device with data calculation, processing and storage capabilities, such as an industrial PC, server, tablet computer, etc. This application embodiment does not limit this.

[0053] Figure 1 A flowchart of a method for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window, as provided in this application embodiment, is shown below. Figure 1 As shown, the method for determining the moisture endpoint includes the following steps:

[0054] Step 101: Acquire multi-source monitoring data during the fluidized bed granulation process through a multi-modal sensing acquisition unit. The multi-source monitoring data includes at least near-infrared spectral data, material moisture data, and particle size data.

[0055] A multimodal sensing acquisition unit refers to an integrated acquisition device that combines multiple sensing modules to simultaneously acquire different types of monitoring data. Its function is to provide comprehensive and real-time raw data support for moisture endpoint determination, ensuring the accuracy and comprehensiveness of subsequent analysis. In this embodiment, the unit specifically includes a near-infrared spectral acquisition module for acquiring near-infrared spectral data, a humidity sensing module for acquiring material moisture data, and a particle size monitoring module for acquiring particle size data. These modules work collaboratively to achieve simultaneous acquisition of multi-dimensional data. Near-infrared spectral data refers to the spectral response information of the material in the near-infrared band (800-2500nm) captured by the near-infrared spectral acquisition module. This includes relevant characteristics such as surface moisture and composition, and is data for analyzing the moisture distribution on the particle surface. Material moisture data refers to the moisture content-related data of the material within the fluidized bed layer acquired by the humidity sensing module, including predicted surface moisture and overall moisture value, reflecting the moisture content level of the material. Particle size data refers to the size-related parameters of particles collected by the particle size monitoring module during the granulation process, including characteristic parameters such as particle size distribution, average particle size, and median particle size. These parameters can be used to help determine the correlation between the granulation process and the moisture endpoint.

[0056] Specifically, the deployment and debugging of the multimodal sensing and acquisition unit are completed first to ensure that each module operates normally and data acquisition is synchronized. Optionally, the near-infrared spectroscopy acquisition module uses a near-infrared spectrometer with a wavelength range of 800-2500nm, a spectral resolution ≤2nm, and a scanning frequency set to 1-5Hz, equipped with a high-temperature resistant and dust-proof fiber optic probe. The fiber optic probe is installed at the observation window above the fluidized bed, with the probe axis at a 45° angle to the surface of the bed material, 15-20cm away from the material surface, ensuring that the probe can comprehensively capture the near-infrared spectral information of the material within the bed without any blind spots. The spectrometer and electronic equipment are connected via a USB industrial bus to transmit the acquired raw near-infrared spectral data in real time, with a transmission delay ≤10ms. Simultaneously, spectral preprocessing software is activated to perform baseline correction and noise reduction on the raw spectral data, eliminating the influence of ambient light, dust, and other interference factors to ensure the accuracy of the near-infrared spectral data. The humidity sensing module uses a high-precision capacitive humidity sensor with a measurement range of 0-100%RH, a measurement accuracy of ±0.5%RH, and a response time ≤1s. The sensor is installed inside the fluidized bed, embedded in the material flow area. The sensor probe is designed to be corrosion-resistant and clogging-resistant to prevent material adhesion from affecting measurement accuracy. This module collects the overall moisture value of the material in real time and works in conjunction with a near-infrared spectroscopy module to assist in generating a predicted surface moisture value. The analog signal output by the sensor is converted into a digital signal by an A / D converter and transmitted to the electronic equipment to ensure the real-time and completeness of the material moisture data. The particle size monitoring module uses an online laser particle size analyzer with a measurement range of 0.1-1000μm and a measurement accuracy of ≤2%, employing the laser scattering principle. The particle size analyzer probe is installed at the fluidized bed outlet at a 30° angle to the material flow direction to ensure real-time monitoring of dynamic changes in particle size and to collect particle size characteristic parameters (including particle size distribution, average particle size, and median particle size). The particle size analyzer establishes real-time communication with the electronic equipment, transmitting the collected particle size data to the electronic equipment in real time at a frequency of 1-5Hz for subsequent dual-index collaborative judgment. At the same time, the data storage function is enabled to record the particle size change data throughout the granulation process, which is convenient for subsequent parameter optimization.

[0057] Start the fluidized bed granulation equipment and preheat it to the preset temperature (determined according to material characteristics, generally 40-60℃). After adding the initial material to the fluidized bed, simultaneously activate the three modules of the multimodal sensing acquisition unit to begin real-time acquisition of near-infrared spectral data, material moisture data, and particle size data. The acquisition process continues until granulation is complete. During the acquisition process, the electronic equipment's memory stores all multi-source monitoring data in real time and marks them with timestamps to ensure data traceability. At the same time, the changing trends of each data point are displayed in real time through the human-machine interface, facilitating operators to monitor the granulation status in real time.

[0058] Step 102: Perform fusion correction processing on the multi-source monitoring data to obtain the fused moisture value.

[0059] Fusion correction processing refers to the process of weighting and anomaly handling of material moisture-related data (predicted surface moisture and overall moisture value) from multi-source monitoring data based on a preset fusion model. This eliminates the limitations of single data points and yields a fused moisture value that accurately represents the actual moisture content of the material, aiming to improve the accuracy and reliability of moisture data. The fused moisture value is a comprehensive moisture index obtained after fusion correction processing, taking into account both the surface and overall moisture state of the material, more realistically reflecting the actual moisture content of the bed material, and providing data for subsequent moisture endpoint determination. The moisture fusion correction model is a mathematical model used to fuse the predicted surface moisture value and the overall moisture value. It involves setting a dynamic fusion weight coefficient, which is dynamically adjusted according to different stages of the granulation process to ensure the accuracy of the fused moisture value at different granulation stages. The dynamic fusion weight coefficient is used in the moisture fusion correction model to adjust the contribution ratio of the predicted surface moisture value and the overall moisture value. Its value changes dynamically with the granulation process (initial, middle, and late stages) to adapt to the characteristics of material moisture distribution at different stages. The predicted surface moisture value refers to the surface moisture content data of the material calculated by the near-infrared spectral acquisition module based on near-infrared spectral data and a preset spectral-moisture prediction model, reflecting the moisture state of the material surface. The overall moisture value refers to the average moisture content data of the material within the fluidized bed directly collected by the humidity sensing module, reflecting the overall moisture level of the material.

[0060] Specifically, after receiving multi-source monitoring data transmitted by the multimodal sensing acquisition unit, the electronic device focuses on the material moisture-related data and initiates the fusion correction processing flow. The specific steps are as follows:

[0061] The predicted surface moisture value output from the near-infrared spectroscopy acquisition module and the overall moisture value output from the humidity sensing module were extracted from the multi-source monitoring data. These two types of data underwent preprocessing. The 3σ criterion was used to remove abnormal data caused by sensor malfunctions and environmental interference, ensuring data accuracy. Simultaneously, both types of moisture data were standardized, converting them into standardized values ​​in the 0-1 range to eliminate the influence of dimensions.

[0062] A moisture fusion correction model was constructed, which uses dynamic fusion weighting coefficients to weight the predicted surface moisture value and the overall moisture value. The dynamic fusion weighting coefficients are dynamically adjusted according to different stages of the granulation process. The specific adjustment logic is as follows: Early granulation stage (0-10 min, material drying stage): At this stage, the material is dry, and the difference between surface moisture and overall moisture is small. The weighting coefficient for the predicted surface moisture value is set to 0.4-0.5, and the weighting coefficient for the overall moisture value is set to 0.5-0.6, ensuring that the overall moisture state dominates the fusion result, which aligns with the moisture distribution characteristics of the material drying stage. Mid-granulation stage (10-30 min, particle formation stage): The material begins to agglomerate and form particles, and the difference between surface moisture and overall moisture is larger. The surface moisture prediction weighting coefficient is set to 0.6-0.7, and the overall moisture value weighting coefficient is set to 0.3-0.4 to highlight the influence of surface moisture on particle formation and ensure that the fusion result can reflect the moisture state of the particle surface. In the later stage of granulation (after 30 minutes, the moisture stabilization stage): the particles gradually take shape, the moisture distribution tends to stabilize, the difference between surface moisture and overall moisture decreases, and the weighting coefficients are restored to the level of the early stage of granulation (surface moisture prediction weighting coefficient 0.4-0.5, overall moisture value weighting coefficient 0.5-0.6) to ensure that the fusion result conforms to the moisture characteristics of the material in the final stable state.

[0063] The weighted calculation based on the moisture fusion correction model yields the fused moisture value, and the calculation formula is as follows: in, Let be the fusion moisture value at time t. The weighting coefficients for the predicted surface moisture values. Let be the predicted surface moisture value at time t. This is the weighting coefficient for the overall moisture value. The total moisture value at time t. After the fused moisture value is calculated, it is stored in real time in the memory of the electronic device and transmitted to the subsequent processing module for subsequent deviation analysis and collaborative determination.

[0064] Step 103: Perform moving window standard deviation analysis on the near-infrared spectral data, construct the spectral fluctuation characteristic function, and determine the moisture distribution state on the particle surface based on the spectral fluctuation characteristic function.

[0065] Moving window standard deviation analysis is a data analysis method that divides near-infrared spectral data into continuous window segments of a preset size, calculates the standard deviation and mean of the spectral absorbance values ​​within each window, and then analyzes the spectral fluctuation patterns. Its function is to capture the local fluctuation characteristics of spectral data, indirectly reflecting the uniformity of moisture distribution on the particle surface. The spectral fluctuation characteristic function is a mathematical function constructed based on the spectral absorbance values ​​and mean absorbance within the moving window. It quantifies the degree of spectral fluctuation, and its changing trend directly corresponds to the change in the distribution state of moisture on the particle surface, serving as a basis for judging whether the moisture on the particle surface is uniform. The particle surface moisture distribution state refers to the uniformity of moisture distribution on the particle surface, including both uniform and non-uniform distributions. This state directly affects granulation quality and is an important reference condition for determining the moisture endpoint. The spectral absorbance value refers to the degree of absorption of near-infrared light by the material at a specific wavelength in the near-infrared spectral data. Its change is related to the moisture content on the material surface and is the basic data for constructing the spectral fluctuation characteristic function. The absorbance mean is the average of all spectral absorbance values ​​within each moving window. It is used for normalization to eliminate the impact of differences in overall absorbance levels between different windows on fluctuation analysis.

[0066] Specifically, after acquiring the preprocessed near-infrared spectral data, the electronic device initiates the moving window standard deviation analysis process, with the following specific steps:

[0067] Determine the size of the moving window. The window size can be adjusted according to the spectral scanning frequency and granulation speed, and is generally set to 5-10 scanning points. Divide the full-band spectral information of the near-infrared spectral data into moving windows according to this window size. Each window contains spectral data of 5-10 consecutive scanning points. The windows move sequentially according to the scanning order, with a moving step size of 1 scanning point, to ensure that all spectral data is covered and no local fluctuation features are missed.

[0068] The spectral absorbance values ​​within each moving window are calculated to obtain the standard deviation (SD) and mean absorbance for each window. The calculation process is performed according to the following formula: ,

[0069] in, denoted as , where is the absorbance value of the i-th scan point within the window, n is the number of scan points within the window, SD is the standard deviation of the spectral absorbance within the window, reflecting the degree of spectral fluctuation within the window, and Mean is the mean of the spectral absorbance within the window, used to eliminate the influence of overall absorbance level differences.

[0070] Based on the SD and Mean within each moving window, a spectral fluctuation characteristic function is constructed, with the function expression as follows:

[0071] in, λ is the spectral wavelength, and t is the granulation time. For time t Standard deviation of the moving window at wavelength. For time t Mean absorbance of the moving window at wavelength For time t The spectral fluctuation characteristic value at a wavelength indicates that the larger the value, the more intense the spectral fluctuation at that wavelength at that moment, and the more uneven the distribution of moisture on the particle surface.

[0072] The surface moisture distribution of particles is determined based on the changing trend of the spectral fluctuation characteristic function. A fluctuation amplitude threshold of 5% is set; when the value continuously decreases and tends to stabilize, and this occurs for 10 consecutive scanning points... When the change in value is ≤5%, the moisture on the particle surface is considered to be uniformly distributed; when When the value fluctuation range is greater than 5%, it is determined that the moisture distribution on the particle surface is uneven. The electronic device records this status information to provide a reference for the endpoint determination.

[0073] Step 104: Based on the fused moisture value and particle size data, a moisture endpoint determination signal is generated through dual-index collaborative determination logic.

[0074] The dual-indicator collaborative judgment logic refers to a logical system that uses both fused moisture value and particle size data as judgment indicators. By constructing a collaborative judgment function, it quantifies the changing patterns of both indicators to determine whether the moisture endpoint has been reached. Its advantage lies in considering both moisture state and granulation process, improving the accuracy of endpoint determination. The rate of change of fused moisture value refers to the amount of change in fused moisture value per unit time, reflecting the dynamic trend of moisture content change. Its magnitude indicates whether the moisture content is stabilizing and is one of the core indicators for determining the moisture endpoint. The calculation formula is dM / dt. The rate of change of particle size characteristic parameters refers to the amount of change in particle size characteristic parameters (such as average particle size) per unit time, reflecting the dynamic process of particle formation. Its magnitude indicates whether the particles are stabilizing and is an important indicator for assisting in determining the moisture endpoint. The calculation formula is dD / dt. The dual-indicator collaborative judgment function is a mathematical function that weights and sums the rates of change of fused moisture value and particle size characteristic parameters. It is used to quantify the degree of collaborative change of the two indicators, and its calculation result is the direct basis for triggering the moisture endpoint judgment signal. The preset judgment conditions refer to the pre-defined threshold conditions used to determine whether the moisture endpoint has been reached. These include the threshold calculated by the collaborative judgment function and the duration threshold. The moisture endpoint judgment signal is triggered only when both threshold requirements are met simultaneously. The moisture endpoint judgment signal is a signal generated by the electronic device when the preset judgment conditions are met, indicating that the moisture endpoint has been reached. It is the trigger signal for initiating endpoint confirmation and subsequent control procedures.

[0075] Specifically, after acquiring the fused moisture value and particle size data, the electronic device initiates a dual-indicator collaborative judgment process, with the following specific steps:

[0076] Calculate the rate of change of the fused moisture content and the rate of change of the particle size characteristic parameters. Set the calculation step size. (Consistent with the sampling frequency, 0.2-1s), the formula for calculating the rate of change of the fused moisture value is: Where dM / dt is the rate of change of the fused moisture value at time t, and M(t) is the fused moisture value at time t. for The fusion moisture value at a given time. The formula for calculating the rate of change of particle size characteristic parameters is: Where dD / dt is the rate of change of the particle size characteristic parameter at time t, and D(t) is the particle size characteristic parameter at time t (using the average particle size). for Particle size characteristic parameters at time t.

[0077] A dual-index collaborative judgment function is constructed, which weights and sums the rates of change of fused moisture value and the rates of change of particle size characteristic parameters. The function expression is as follows:

[0078]

[0079] in, The result of the collaborative decision function calculation at time t. , These are weighting coefficients. The value range is 0.6-0.7. The value ranges from 0.3 to 0.4 and can be adjusted according to the material characteristics. In this embodiment, the default value is used. =0.65, =0.35, this weight allocation highlights the dominant role of the fusion moisture value in the determination of the endpoint, while also taking into account the auxiliary reference value of particle size.

[0080] Set a preset judgment condition. The absolute value of the result calculated by the collaborative judgment function is always ≤0.05 and the duration is ≥30s. The electronic device compares the value of G(t) with the preset judgment condition in real time and records the duration for which the value of G(t) satisfies |G(t)|≤0.05.

[0081] When the absolute value of G(t) remains ≤0.05 for 30 seconds, the electronic device triggers the moisture endpoint determination signal and transmits the signal to the endpoint determination module to initiate the subsequent endpoint confirmation process. If the G(t) value does not meet the preset conditions, monitoring continues until all conditions are met and the endpoint signal is triggered.

[0082] Step 105: Based on the moisture endpoint determination signal, determine the moisture endpoint of the fluidized bed granulation process.

[0083] Moisture endpoint determination refers to the process of confirming that the fluidized bed granulation process has reached the moisture endpoint based on the moisture endpoint determination signal, combined with the moisture distribution on the particle surface and verification using a two-dimensional dynamic characteristic spectrum. This ensures the accuracy and reliability of the endpoint determination and provides a basis for subsequent granulation control. The two-dimensional dynamic characteristic spectrum is a visual graph constructed with granulation time as the horizontal axis and fused moisture value and particle size characteristic parameters as the vertical axes. It can intuitively present the dynamic evolution trend of moisture and particle size during granulation and is an important tool to assist in confirming the moisture endpoint. The moisture endpoint determination region refers to the reasonable range defined in the two-dimensional dynamic characteristic spectrum that meets the moisture and particle size requirements of qualified products. It serves as a visual reference boundary for determining whether the endpoint has been reached. The preset residence time requirement refers to the minimum continuous residence time required after the fused moisture value and particle size characteristic parameters simultaneously enter the moisture endpoint determination region, used to avoid misjudgments caused by data fluctuations. The granulation endpoint confirmation signal is the signal generated by the electronic equipment after verification using the two-dimensional dynamic characteristic spectrum, confirming that the granulation endpoint has been reached. It is the formal signal to end granulation or initiate subsequent control.

[0084] Specifically, after receiving the moisture endpoint determination signal, the electronic device initiates the moisture endpoint confirmation process, with the following steps:

[0085] A two-dimensional dynamic feature spectrum is constructed. With granulation time as the horizontal axis and the fusion moisture value and particle size characteristic parameters as the vertical axes, the change curves of the two are plotted in real time to form a two-dimensional dynamic feature spectrum. This spectrum intuitively presents the dynamic evolution trend of moisture and particle size during the granulation process, facilitating real-time monitoring of the granulation status by operators.

[0086] Define the moisture endpoint determination region. Based on the material characteristics and granulation quality requirements, define a reasonable moisture endpoint determination region in the two-dimensional dynamic characteristic spectrum. This region is the acceptable range of the blended moisture value and particle size characteristic parameters (e.g., blended moisture value of 8%-12%RH, average particle size of 100-200μm). The range of the region can be adjusted according to actual production needs to ensure that the determination region can accurately correspond to the parameter range of the qualified finished product.

[0087] Endpoint confirmation is performed. The electronic device monitors in real time whether the fused moisture value and particle size characteristic parameters simultaneously enter the preset moisture endpoint determination area and records the residence time. When both enter the determination area simultaneously and the residence time is ≥20s, combined with the judgment result of the particle surface moisture distribution status in step S103 (the particle surface moisture tends to be uniformly distributed), the electronic device issues a granulation endpoint confirmation signal, officially completing the determination of the moisture endpoint in the fluidized bed granulation process; if both do not enter the determination area simultaneously, or the residence time is insufficient, or the particle surface moisture distribution is uneven, the endpoint confirmation is delayed, and monitoring continues until all conditions are met.

[0088] After the endpoint is confirmed, the human-machine interface displays the endpoint determination results (including endpoint time, fusion moisture value, and particle size parameters). At the same time, the electronic device stores the endpoint-related data to facilitate subsequent production traceability and parameter optimization.

[0089] Furthermore, the moisture endpoint determination method provided in this application embodiment also includes an adaptive control step for the granulation process: the electronic device generates a granulation process control command to adaptively adjust the feeding rate during the granulation process. Specifically, after the moisture endpoint determination signal is triggered, the feeding rate is gradually reduced by 20%-30%, and the granulation equipment is kept running for 5-10 minutes to ensure stable granule formation and avoid excessively wet granules or uneven particle size due to excessive feeding. After the adjustment is completed, the granulation equipment and the multimodal sensor acquisition unit can be turned off according to actual production needs to stop data acquisition.

[0090] After granulation, the finished granules are sampled and tested to verify the accuracy of the moisture endpoint determination. If the granule moisture content and particle size distribution meet the preset quality requirements, the parameters of this granulation (dynamic fusion weighting coefficient, weighting coefficient, preset judgment conditions, etc.) are stored as a high-quality template for subsequent granulation production of materials of the same specifications. If the test results do not meet the requirements, the reasons for the deviation are analyzed, and relevant parameters (such as moving window size, weighting coefficient, endpoint determination area range, etc.) are adjusted to optimize the judgment logic and improve the accuracy of subsequent endpoint determination.

[0091] The above steps, verified by visual graphs, further improve the accuracy of endpoint determination, avoid misjudgments caused by fluctuations in a single indicator, and ensure the stability of granulation quality.

[0092] Figure 2 This diagram illustrates the structure of a moisture endpoint determination device for a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window, as provided in an embodiment of this application. This device is configured to execute the moisture endpoint determination method provided in the above embodiment, possessing the corresponding functional modules and beneficial effects for executing the method. Figure 2 As shown, the device specifically includes:

[0093] Data acquisition module 201 is used to acquire multi-source monitoring data in the fluidized bed granulation process through a multi-modal sensing acquisition unit. The multi-source monitoring data includes at least near-infrared spectral data, material moisture data, and particle size data.

[0094] The fusion correction module 202 is used to perform fusion correction processing on the multi-source monitoring data to obtain a fused moisture value;

[0095] The deviation analysis module 203 is used to perform moving window standard deviation analysis on the near-infrared spectral data, construct a spectral fluctuation characteristic function, and determine the moisture distribution state on the particle surface based on the spectral fluctuation characteristic function.

[0096] The collaborative determination module 204 is used to generate a moisture endpoint determination signal based on the fused moisture value and particle size data through dual-index collaborative determination logic.

[0097] The endpoint determination module 205 is used to determine the moisture endpoint in the fluidized bed granulation process based on the moisture endpoint determination signal.

[0098] The aforementioned device, through the collaborative work of its various modules, achieves synchronous acquisition, fusion correction, spectral analysis, collaborative judgment, and endpoint output of multi-source data. It can ensure the integrity of the installation package file while providing a structured carrier for unified configuration logic. This allows the installation package to directly query interface configuration based on the configuration bytecode file during runtime, without the need for reflection mechanisms. This reduces the resource consumption of interface configuration queries and improves the efficiency of interface configuration queries.

[0099] In one possible embodiment, the data acquisition module 201 is specifically configured as follows:

[0100] The multimodal sensing and acquisition unit includes a near-infrared spectral acquisition module for acquiring near-infrared spectral data, a humidity sensing module for acquiring material moisture data, and a particle size monitoring module for acquiring particle size data.

[0101] In one possible embodiment, the fusion correction module 202 is specifically configured as follows:

[0102] Receive the surface moisture prediction value output by the near-infrared spectroscopy acquisition module and the overall moisture value output by the humidity sensing module;

[0103] A moisture fusion correction model is constructed based on the predicted surface moisture value and the overall moisture value. The moisture fusion correction model uses dynamic fusion weight coefficients to weight the predicted surface moisture value and the overall moisture value to obtain the fused moisture value.

[0104] The dynamic fusion weighting coefficient is dynamically adjusted according to different stages of the granulation process.

[0105] In one possible embodiment, the deviation analysis module 203 is specifically configured as follows:

[0106] Moving window standard deviation analysis was performed on the full-band spectral information of near-infrared spectral data to obtain spectral absorbance values ​​and mean absorbance values;

[0107] Based on the spectral absorbance value and the average absorbance value within the moving window, a spectral fluctuation characteristic function is constructed;

[0108] Based on the changing trend of the spectral fluctuation characteristic function, it is determined whether the moisture on the particle surface tends to be uniformly distributed.

[0109] In one possible embodiment, the collaborative determination module 204 is specifically configured as follows:

[0110] Obtain the rate of change of fusion moisture value and the rate of change of particle size characteristic parameters;

[0111] A dual-index collaborative judgment function is constructed based on the rate of change of the fusion moisture value and the rate of change of the particle size characteristic parameter. The judgment function performs a weighted summation of the rate of change of the fusion moisture value and the rate of change of the particle size characteristic parameter.

[0112] When the calculation result of the determination function continuously meets the preset determination conditions, the moisture endpoint determination signal is triggered.

[0113] In one possible embodiment, after the endpoint determination module 205, it is further configured as follows:

[0114] A two-dimensional dynamic feature spectrum was constructed with granulation time as the horizontal axis and fusion moisture value and particle size characteristic parameters as the vertical axes, respectively.

[0115] A moisture endpoint determination region is defined in the two-dimensional dynamic feature spectrum. When the fused moisture value and particle size characteristic parameters simultaneously enter the determination region and meet the preset residence time requirement, a granulation endpoint confirmation signal is issued.

[0116] In one possible embodiment, after the endpoint determination module 205, it is further configured as follows:

[0117] After receiving the moisture endpoint determination signal, it generates a granulation process control command to adaptively adjust the feed rate during the granulation process.

[0118] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 3 As shown, the electronic device includes a processor 301, a memory 302, an input device 303, and an output device 304.

[0119] The number of processors 301 can be one or more. Figure 3 Taking a processor 301 as an example; the processor 301, memory 302, input device 303, and output device 304 can be connected via a bus or other means. Figure 3 Taking a bus connection as an example, the memory 302, as a computer-readable storage medium, can be configured to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the moisture endpoint determination method in this embodiment. The processor 301 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 302, thereby realizing the aforementioned moisture endpoint determination method for the fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of the moving window. The input device 303 can be configured to receive input digital or character information and generate key signal inputs related to user settings and function control of the electronic device. The output device 304 may include a display screen or other display device for displaying monitoring data, determination results, and other information during the granulation process.

[0120] This application also provides a non-volatile storage medium containing computer-executable instructions. When executed by a computer processor, the computer-executable instructions are configured to execute a method for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with moving window standard deviation. The method includes: acquiring multi-source monitoring data during the fluidized bed granulation process through a multi-modal sensing acquisition unit; the multi-source monitoring data including at least near-infrared spectral data, material moisture data, and particle size data; performing fusion correction processing on the multi-source monitoring data to obtain a fused moisture value; performing moving window standard deviation analysis on the near-infrared spectral data to construct a spectral fluctuation characteristic function, and determining the moisture distribution state on the particle surface based on the spectral fluctuation characteristic function; generating a moisture endpoint determination signal based on the fused moisture value and particle size data through a dual-index collaborative determination logic; and determining the moisture endpoint in the fluidized bed granulation process based on the moisture endpoint determination signal.

[0121] It is worth noting that in the embodiments of the moisture endpoint determination device described above, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not configured to limit the protection scope of the embodiments of this application. The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0122] The same or similar labels correspond to the same or similar parts;

[0123] The terms used to describe positional relationships in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent.

[0124] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. For those skilled in the art, other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all implementation methods here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the claims of the present invention.

Claims

1. A method for determining the moisture endpoint in a fluid bed granulation process based on near infrared spectroscopy combined with moving window standard deviation, characterized in that, The determination method includes: Multi-source monitoring data during the fluidized bed granulation process is acquired through a multi-modal sensing acquisition unit. The multi-source monitoring data includes at least near-infrared spectral data, material moisture data, and particle size data. The multi-source monitoring data are fused and corrected to obtain a fused moisture value; A moving window standard deviation analysis was performed on the near-infrared spectral data to construct a spectral fluctuation characteristic function, and the surface moisture distribution of the particles was determined based on the spectral fluctuation characteristic function. Based on the fused moisture value and particle size data, a moisture endpoint determination signal is generated through a dual-index collaborative determination logic. Based on the moisture endpoint determination signal, the moisture endpoint of the fluidized bed granulation process is determined.

2. The determination method according to claim 1, characterized by The multimodal sensing and acquisition unit includes a near-infrared spectral acquisition module for acquiring near-infrared spectral data, a humidity sensing module for acquiring material moisture data, and a particle size monitoring module for acquiring particle size data.

3. The determination method according to claim 1, characterized by, The process of fusing and correcting the multi-source monitoring data to obtain the fused moisture value includes: Receive the surface moisture prediction value output by the near-infrared spectroscopy acquisition module and the overall moisture value output by the humidity sensing module; A moisture fusion correction model is constructed based on the predicted surface moisture value and the overall moisture value. The moisture fusion correction model uses dynamic fusion weight coefficients to weight the predicted surface moisture value and the overall moisture value to obtain the fused moisture value. The dynamic fusion weighting coefficient is dynamically adjusted according to different stages of the granulation process.

4. The determination method according to claim 1, characterized by, The step of performing moving window standard deviation analysis on the near-infrared spectral data to construct a spectral fluctuation characteristic function includes: Moving window standard deviation analysis was performed on the full-band spectral information of near-infrared spectral data to obtain spectral absorbance values ​​and mean absorbance values; Based on the spectral absorbance value and the average absorbance value within the moving window, a spectral fluctuation characteristic function is constructed; Based on the changing trend of the spectral fluctuation characteristic function, it is determined whether the moisture on the particle surface tends to be uniformly distributed.

5. The determination method according to claim 1, characterized by, The step of generating a moisture endpoint determination signal based on the fused moisture value and particle size data, through a dual-indicator collaborative determination logic, includes: Obtain the rate of change of fusion moisture value and the rate of change of particle size characteristic parameters; A dual-index collaborative judgment function is constructed based on the rate of change of the fusion moisture value and the rate of change of the particle size characteristic parameter. The judgment function performs a weighted summation of the rate of change of the fusion moisture value and the rate of change of the particle size characteristic parameter. When the calculation result of the determination function continuously meets the preset determination conditions, the moisture endpoint determination signal is triggered.

6. The determination method according to claim 5, characterized in that, The method further includes: A two-dimensional dynamic feature spectrum was constructed with granulation time as the horizontal axis and fusion moisture value and particle size characteristic parameters as the vertical axes, respectively. A moisture endpoint determination region is defined in the two-dimensional dynamic feature spectrum. When the fused moisture value and particle size characteristic parameters simultaneously enter the determination region and meet the preset residence time requirement, a granulation endpoint confirmation signal is issued.

7. The method of claim 1, wherein, The method further includes: after receiving the moisture endpoint determination signal, generating a granulation process control command to adaptively adjust the feed rate during the granulation process.

8. A fluid bed granulation process moisture endpoint determination device based on near infrared spectroscopy combined with moving window standard deviation, characterized in that, The device includes: The data acquisition module is used to acquire multi-source monitoring data during the fluidized bed granulation process through a multi-modal sensing acquisition unit. The multi-source monitoring data includes at least near-infrared spectral data, material moisture data, and particle size data. The fusion correction module is used to perform fusion correction processing on the multi-source monitoring data to obtain a fused moisture value; The deviation analysis module is used to perform moving window standard deviation analysis on the near-infrared spectral data, construct a spectral fluctuation characteristic function, and determine the moisture distribution state on the particle surface based on the spectral fluctuation characteristic function. The collaborative determination module is used to generate a moisture endpoint determination signal based on the fused moisture value and particle size data through dual-index collaborative determination logic. The endpoint determination module is used to determine the moisture endpoint in the fluidized bed granulation process based on the moisture endpoint determination signal.

9. An electronic device, comprising: The electronic device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the method for determining the moisture endpoint in the fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of the moving window as described in any one of claims 1-7.

10. A computer storage medium, characterized in that, The computer storage medium stores computer instructions, which, when invoked, are used to execute the method for determining the moisture endpoint in a fluidized bed granulation process based on near-infrared spectroscopy combined with the standard deviation of a moving window as described in any one of claims 1-7.

Images