Fertilizer grade magnesium sulfate monohydrate chemical process control system based on digital twinning

Abstract

The application discloses a fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twinning. The system obtains double-activity parameters of the surface and the inside through terahertz internal perspective, establishes a core-shell decoupling model, and expands the detection dimension from the surface to the whole. At the same time, the core pre-release index is generated through the initial thermal-acoustic joint monitoring, the activity transition risk is predicted in advance, the control strategy is changed from passive response to active pre-adjustment, and the later loss of control is avoided. The double-zone reconstruction module establishes the heterogeneous coupling kinetics framework of the surface dissolution and the core reaction, dynamically switches the control domain and the model parameters, and solves the defect that the single-zone model cannot represent the heterogeneous reaction characteristics. The control verification module realizes the continuous calibration and self-optimization of the model parameters through online phase detection and feedback correction, and ensures the model accuracy in long-term operation.

Description

Technical Field

[0001] This application relates to the field of processing technology, and in particular to a control system for the formation process of fertilizer-grade magnesium sulfate monohydrate based on digital twins. Background Technology

[0002] With the in-depth development of intelligent manufacturing technology in the process industry, digital twin technology has gradually penetrated from macroscopic equipment monitoring to microscopic reaction mechanism modeling, forming a high-precision control system for complex physicochemical processes in the fields of hydrometallurgy and inorganic salt chemical engineering. Among them, the formation process of fertilizer-grade magnesium sulfate monohydrate, as a typical gas-liquid-solid three-phase strongly exothermic reaction system, makes the feedforward control of raw material quality a key link in determining product quality and production safety. In particular, the online characterization technology of the activity of the core raw material, light-burned magnesium oxide (active MgO), has become a research hotspot and technical challenge in the digital transformation of this field.

[0003] However, existing technologies generally employ near-infrared spectroscopy (NIR) for rapid detection and activity pre-assessment of light-calcined magnesia raw materials. This technology can only penetrate the surface structure of particles up to 50–100 μm to obtain vibrational absorption information of surface functional groups. Because light-calcined magnesia exhibits a typical "surface overburning – core underburning" core-shell structure during calcination, the dense periclase lattice formed on the particle surface displays low activity characteristics, while the internal core retains magnesium hydroxide or magnesium carbonate precursors due to thermal conduction lag and possesses high potential activity. This results in a significant deviation between the "standard activity" (α≈1.0) detected by NIR and the actual overall activity of the particles. This asymmetry in surface-core activity information causes the kinetic parameters corrected by the digital twin system based on pseudo-standard activity to lag significantly in the early stages of the reaction. When the highly active MgO core is suddenly exposed after 15–40 minutes of reaction, the system still maintains a medium-rate acid addition strategy and cannot predict the exponential jump in reaction rate (equivalent activity jumps from 1.0 to 1.4). This ultimately leads to uncontrollable boiling, local overheating, and the formation of insoluble anhydrous magnesium sulfate that clogs the pipes in the later stages of the reaction, as well as excessive water-insoluble matter in the product (up to 2% or more, far exceeding the fertilizer grade standard of <0.5%), resulting in unplanned shutdowns and batch quality incidents. Summary of the Invention

[0004] The main objective of this application is to provide a digital twin-based control system for the formation process of fertilizer-grade magnesium sulfate monohydrate, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, this application provides the following technical solution: A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin, the system includes a raw material pre-screening module, a terahertz scanning module, a transition early warning module, a dual-zone reconstruction module and a control verification module; The raw material pre-screening module is configured at the feed port of the light-burned magnesium oxide raw material silo. Based on the results of spectral detection, visual recognition and particle size analysis, it identifies and removes extreme particles with severe overburning or complete underburning, and sends medium core-shell structure particles into the terahertz perspective module. The terahertz perspective module is connected to the multimodal pre-screening module. It obtains dielectric response characteristics by penetrating the interior of the particles through terahertz electromagnetic waves, analyzes the surface activity coefficient and the core activity coefficient, and establishes a core-shell activity decoupling model. The transition early warning module is installed inside the reactor. During the initial stage of the reaction, it monitors the micro-thermal changes and ultrasonic pulse response during the dissolution of the overburned layer on the surface, generates the core pre-release index, and predicts the activity transition time by combining the surface activity coefficient and the core activity coefficient. The dual-zone reconstruction module is connected to the terahertz vision module and the early warning module respectively. It receives the surface activity coefficient, the core activity coefficient and the core pre-release index, and constructs a dual-zone coupled dynamic model of the surface dissolution zone and the core reaction zone. When the core is predicted to be exposed, the control equation and model parameters are dynamically switched. The control verification module is connected to the dual-zone reconstruction module. Based on the output of the dual-zone coupled dynamic model, it executes a segmented pre-adjustment control strategy and performs online verification of the phase state of the discharged crystal. The verification results are fed back to the terahertz imaging module to correct the detection parameters.

[0006] Preferably, the terahertz fluoroscopy module includes a terahertz tomography unit and a dual-zone active decoupling unit; The terahertz tomography unit is located in the light-burned magnesium oxide particle transport channel. It emits terahertz electromagnetic waves to penetrate the particles and collect time-domain spectral signals. The internal dielectric constant distribution is analyzed by the tomographic reconstruction algorithm, and the characteristic absorption peaks of magnesium hydroxide or magnesium carbonate in the core region are extracted to determine the core-shell interface position and internal structural features. The dual-zone active decoupling unit receives the internal structural features output by the tomographic scanning unit and the surface functional group information output by the upstream surface detection unit, establishes an active mapping model between the overburned surface layer and the underburned core region, calculates the surface activity coefficient and the core activity coefficient, constructs a core-shell active decoupling model, and outputs dual parameters to the downstream module.

[0007] Preferably, the terahertz tomography unit includes a terahertz spectroscopy unit and a tomography reconstruction unit; The terahertz spectral unit is equipped with a transmission detection component and a pulse emission source. It emits terahertz electromagnetic waves that penetrate the entire scale of lightly calcined magnesium oxide particles, collects the time-domain electric field waveform after being modulated by different internal orientation regions, records the coordinated response characteristics of amplitude attenuation and phase delay, generates raw spectral data containing internal dielectric information, and transmits it to the tomographic reconstruction unit. The tomographic reconstruction unit is connected to the terahertz spectral unit to receive the raw spectral data. Based on the filtered back-projection tomographic reconstruction algorithm, it calculates the three-dimensional distribution of the dielectric constant inside the particle. It identifies and extracts the residual magnesium hydroxide or magnesium carbonate fingerprint lines in the core region through characteristic absorption peaks, locates the core-shell interface and internal heterogeneous structural features, and outputs structural parameters to the dual-zone active decoupling unit.

[0008] Preferably, the dual-region activity decoupling unit includes a feature fusion unit and an activity inversion unit; The feature fusion unit is connected to the terahertz tomography unit and the surface detection unit. It receives the internal dielectric constant distribution and the absorption spectrum of surface functional groups, establishes a spatiotemporal alignment mapping of the core and shell features, constructs a fusion dataset related to the core and shell positions and surface activity, and outputs it downstream. The activity inversion unit is connected to the feature fusion unit. Based on the fusion dataset, a dielectric response-activity mapping between the overburned surface layer and the underburned core region is constructed. The surface activity coefficient and the core activity coefficient are inverted through the difference in dielectric constant. A core-shell activity decoupling model is established and dual activity parameters are output downstream.

[0009] Preferably, the transition early warning module includes a thermal acoustic monitoring unit and a transition prediction unit; The thermoacoustic monitoring unit is configured on the inner wall of the reactor and in the stirring shaft area. It integrates a highly sensitive micro-heat detection component and a broadband ultrasonic pulse transceiver component. In the initial stage of the reaction, it synchronously collects the weak heat flow changes and slurry sound velocity fluctuations during the dissolution process of the surface overburnt layer. It constructs a time-series dataset containing the heat of dissolution phase change and acoustic propagation characteristics and outputs it downstream. The transition prediction unit is connected to the thermoacoustic monitoring unit and the upstream terahertz perspective module, respectively. It receives time-series datasets and constructs a core pre-release index calculation model by combining the surface activity coefficient and the core activity coefficient. It analyzes the collapse dynamic trend of the surface overburnt layer, predicts the exposure time of the core high-activity zone and the equivalent activity transition amplitude, and outputs transition warning information to the dual-zone reconstruction module.

[0010] Preferably, the thermoacoustic monitoring unit includes a heat detection unit and an acoustic wave detection unit; The heat detection unit is arranged in the area near the liquid surface on the inner wall of the reactor. It integrates a high-sensitivity micro-heat sensor array to continuously collect the weak heat flow changes and phase change heat effects during the dissolution process of the surface overburnt layer in the initial stage of the reaction. It extracts the thermal characteristic information of dissolution kinetics and outputs it synchronously with the acoustic wave detection unit. The acoustic detection unit is located near the impeller area of ​​the stirring shaft. It is equipped with a broadband ultrasonic pulse transceiver array, which transmits detection pulses to the reaction slurry and receives the propagation echo modulated by the particle interface. It monitors the sound velocity fluctuation and energy attenuation characteristics, extracts the acoustic response of surface layer collapse, and merges it with thermal characteristics for downstream output.

[0011] Preferably, the transition prediction unit includes an index construction unit and a trend prediction unit; The index construction unit is connected to the thermoacoustic monitoring unit and the dual-zone activity decoupling unit respectively. It receives the thermoacoustic fusion dataset and the surface-shell activity coefficients, establishes the correlation mapping between surface dissolution and core exposure based on the core-shell interface, constructs the core pre-release index model, and outputs it to the trend prediction unit. The trend prediction unit is connected to the index construction unit and the dual-zone reconstruction module respectively. Based on the core pre-release index, it quantitatively analyzes the dynamic trend of surface overburning layer dissolution and collapse, calculates the exposure time and equivalent activity transition amplitude of the core high-activity zone, generates transition early warning information containing transition time and amplitude, and transmits it to the dual-zone reconstruction module.

[0012] Preferably, the dual-zone reconstruction module includes a dual-zone modeling unit and a dynamic switching unit; The dual-zone modeling unit is connected to the terahertz perspective module and the transition early warning module respectively. It receives the surface activity coefficient, core activity coefficient and core pre-release index. Based on the heterogeneous divide-and-conquer strategy, it constructs a coupled framework of surface overburnt layer shrinkage and dissolution and core underburnt zone reaction kinetics. It analyzes the evolution law of interface mass transfer resistance and core-shell exposed area in real time, generates the dual-zone heterogeneous model skeleton under the current working condition and outputs it to the dynamic switching unit. The dynamic switching unit is connected to the dual-zone modeling unit and the closed-loop control module respectively. It receives the dual-zone heterogeneous model skeleton and core pre-release index, monitors the core pre-release index threshold determination result, triggers smooth switching of the control domain when it predicts that the core is about to be exposed, performs hot-plugging of model structure and online reloading of reaction rate constant and interface mass transfer coefficient, and outputs the dual-zone coupling control parameter set after switching downstream.

[0013] Preferably, the dual-zone modeling unit includes a partition modeling unit and a coupled analytical unit; The partitioned modeling unit is connected to the terahertz perspective module, receives the surface activity coefficient and the core activity coefficient, constructs the control equation for the nucleus shrinkage and dissolution of the overburned surface layer and the first-order reaction kinetic equation for the underburned core region, independently analyzes the reaction rate characteristics and phase transition boundary conditions of the two regions, generates heterogeneous sub-models of the surface region and the core region and outputs them to the coupled analytical unit. The coupled analytical unit is connected to the partitioned modeling unit and the transition early warning module. It receives the heterogeneous sub-model and the core pre-release index, establishes the correlation between the interfacial mass transfer resistance of the surface dissolution zone and the core reaction zone, analyzes the migration trajectory of the core-shell interface and the evolution law of the core exposed area in real time, constructs a dual-zone dynamic coupling skeleton, and outputs the model structure to the dynamic switching unit.

[0014] Preferably, the dynamic switching unit includes a threshold monitoring unit and a model switching unit; The threshold monitoring unit establishes a judgment logic based on the critical threshold of the core pre-release index. It continuously monitors the index change trajectory and growth rate through a sliding window algorithm. When it predicts that the core high-activity area is about to be exposed, it generates a switching trigger command to achieve a smooth transition prediction from the surface dissolution control domain to the dual-zone coupling control domain. The model switching unit receives the switching trigger command and the dual-zone heterogeneous model skeleton, performs hot-plugging of the digital twin model structure and smooth migration of the reaction control domain, reloads the reaction rate constant and interface mass transfer coefficient of the surface dissolution zone and the core reaction zone online, reconstructs the dual-zone coupled control parameter set and transmits it to the closed-loop control module.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This system obtains dual-activity parameters (surface and shell) through terahertz internal imaging, establishing a core-shell decoupled model and expanding the detection dimension from the surface to the whole. Simultaneously, it generates a core pre-release index through thermoacoustic joint monitoring in the early stages of the reaction, predicting the risk of activity transitions in advance and shifting the control strategy from passive response to active pre-adjustment to avoid later loss of control.

[0016] 2. The dual-zone reconstruction module establishes a heterogeneous coupling kinetic framework for surface dissolution and core reactions, dynamically switching control domains and model parameters to address the limitation of single-zone models in characterizing heterogeneous reaction properties. The control verification module achieves continuous calibration and self-optimization of model parameters through online phase detection and feedback correction, ensuring the accuracy of the model during long-term operation. Attached Figure Description

[0017] Figure 1 This is the system flowchart for this application. Detailed Implementation

[0018] 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 a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0019] The terms "first," "second," and "third" in this application are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first," "second," or "third" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified. All directional indications (such as up, down, left, right, front, back, etc.) in the embodiments of this application are only used to explain the relative positional relationships and movements between components in a specific orientation (as shown in the figures). If the specific orientation changes, the directional indications also change accordingly. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0020] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0021] Example 1: Please refer to Figure 1 A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twins, the system includes a raw material pre-screening module, a terahertz scanning module, a transition early warning module, a dual-zone reconstruction module and a control verification module; The raw material pre-screening module is configured at the feed port of the light-burned magnesium oxide raw material silo. Based on the results of spectral detection, visual recognition and particle size analysis, it identifies and removes extreme particles with severe overburning or complete underburning, and sends medium core-shell structure particles into the terahertz perspective module. The terahertz perspective module is connected to the multimodal pre-screening module. It obtains dielectric response characteristics by penetrating the interior of the particles through terahertz electromagnetic waves, analyzes the surface activity coefficient and the core activity coefficient, and establishes a core-shell activity decoupling model. The transition early warning module is installed inside the reactor. During the initial stage of the reaction, it monitors the micro-thermal changes and ultrasonic pulse response during the dissolution of the overburned layer on the surface, generates the core pre-release index, and predicts the activity transition time by combining the surface activity coefficient and the core activity coefficient. The dual-zone reconstruction module is connected to the terahertz vision module and the early warning module respectively. It receives the surface activity coefficient, the core activity coefficient and the core pre-release index, and constructs a dual-zone coupled dynamic model of the surface dissolution zone and the core reaction zone. When the core is predicted to be exposed, the control equation and model parameters are dynamically switched. The control verification module is connected to the dual-zone reconstruction module. Based on the output of the dual-zone coupled dynamic model, it executes a segmented pre-adjustment control strategy and performs online verification of the phase state of the discharged crystal. The verification results are fed back to the terahertz imaging module to correct the detection parameters.

[0022] In this embodiment: the raw material pre-screening module is configured at the discharge port of the light-burned magnesium oxide raw material silo to perform primary sorting of the incoming particles, remove extreme abnormal particles with severely overburned or completely underburned surfaces, and send particles with medium core-shell structure to the downstream detection unit to avoid interference from extreme samples to precision detection equipment and provide a homogenized sample basis for internal activity fluoroscopy.

[0023] The terahertz imaging module utilizes the characteristic of terahertz electromagnetic waves to penetrate the entire particle, collects internal time-domain spectral signals and reconstructs the dielectric constant distribution, identifies the characteristic absorption peaks of residual precursors in the core region, analyzes the surface activity coefficient and the core activity coefficient, and establishes a core-shell activity decoupling model. This overcomes the limitation of traditional spectroscopy, which can only detect the surface, and obtains the true activity information of the entire particle.

[0024] The transition early warning module is placed inside the reactor. In the initial stage of the reaction, it simultaneously collects the micro-thermal changes and ultrasonic pulse response during the dissolution process of the overburned layer on the surface, generates the core pre-release index, and predicts the activity transition moment by combining the surface activity coefficient and the core activity coefficient. It identifies the risk that the core high-activity region will be exposed and avoids sudden temperature rise and pressure increase in the later stage of the reaction.

[0025] The dual-zone reconstruction module receives the surface activity coefficient, core activity coefficient, and core pre-release index, and constructs the kinetic equations for the surface dissolution zone and the core reaction zone, respectively. It establishes a heterogeneous coupling framework, analyzes the evolution of interfacial mass transfer resistance and core-shell exposure area in real time, and dynamically switches the control equations and model parameters when the core is about to be exposed, so that the digital twin model changes from the single-zone homogeneous assumption to dual-zone heterogeneous reconstruction.

[0026] The control verification module executes a segmented pre-adjustment control strategy based on the output of the dual-zone coupled kinetic model. It implements differentiated acid addition rate control at different stages of the reaction and verifies the phase state of the discharged crystals online. The verification results are fed back to the detection front end to correct the model parameters, thereby achieving the matching between the control strategy and the actual activity of the material and ensuring stable product quality.

[0027] Compared to existing technologies that employ near-infrared surface detection and single-region homogeneous reaction models, leading to misjudgments of core-shell structures and late-stage reaction runaway, this system constructs a complete digital twin closed loop from raw material inlet to reaction endpoint through the synergistic operation of five modules: raw material pre-screening, terahertz imaging, transition early warning, dual-region reconstruction, and control verification. Existing technologies, due to limitations in detection depth, cannot identify the internal activity gradient of particles, resulting in a lag in control strategies when the core highly active region is suddenly exposed in the mid-to-late stages of the reaction, leading to violent boiling and phase degradation. This system obtains surface and internal dual activity parameters through terahertz internal imaging, establishing a core-shell decoupled model and expanding the detection dimension from the surface to the entire system. Simultaneously, by generating a core pre-release index through thermoacoustic joint monitoring in the early stages of the reaction, it anticipates the risk of activity transitions, shifting the control strategy from passive response to active pre-adjustment to avoid late-stage runaway. The dual-region reconstruction module establishes a heterogeneous coupling kinetic framework for surface dissolution and core reaction, dynamically switching control domains and model parameters to overcome the limitation of single-region models in characterizing heterogeneous reaction characteristics. The control and verification module achieves continuous calibration and self-optimization of model parameters through online phase detection and feedback correction, ensuring the accuracy of the model during long-term operation. This system eliminates the risk of reaction runaway caused by misjudgment of core-shell structure, avoids the formation of anhydrous magnesium sulfate and pipeline blockage, and substantially improves the stability and controllability of the fertilizer-grade magnesium sulfate monohydrate production process.

[0028] Example 2: Please refer to Figure 1 The terahertz imaging module includes a terahertz tomography unit and a dual-zone active decoupling unit. The terahertz tomography unit is located in the light-burned magnesium oxide particle transport channel. It emits terahertz electromagnetic waves to penetrate the particles and collect time-domain spectral signals. The internal dielectric constant distribution is analyzed by the tomographic reconstruction algorithm, and the characteristic absorption peaks of magnesium hydroxide or magnesium carbonate in the core region are extracted to determine the core-shell interface position and internal structural features. The dual-zone active decoupling unit receives the internal structural features output by the tomographic scanning unit and the surface functional group information output by the upstream surface detection unit, establishes an active mapping model between the overburned surface layer and the underburned core region, calculates the surface activity coefficient and the core activity coefficient, constructs a core-shell active decoupling model, and outputs dual parameters to the downstream module.

[0029] In this embodiment, the terahertz tomography unit is located in the light-burned magnesium oxide particle transport channel. Utilizing the penetrating characteristics of terahertz electromagnetic waves, it completely penetrates the particle scale to acquire internal time-domain spectral signals. Through tomographic reconstruction algorithms, the three-dimensional distribution of the dielectric constant within the particles is calculated, accurately identifying the characteristic absorption peaks of residual magnesium hydroxide or magnesium carbonate in the core region, and determining the core-shell interface location and internal heterogeneous structural features. This unit completes the non-destructive extraction of activity information within the particles, overcoming the physical limitation of traditional near-infrared spectroscopy, which can only detect surface micrometer-level depths. It achieves the goal of visualizing the core-shell structure of light-burned magnesium oxide, providing a true internal structural basis for subsequent activity decoupling.

[0030] The dual-zone activity decoupling unit receives internal structural features and upstream surface detection information, establishes a dielectric response activity mapping model between the overburned surface layer and the underburned core region, calculates the surface activity coefficient and core activity coefficient through dielectric constant difference inversion, constructs a core-shell activity decoupling model, and outputs two parameters downstream. This unit completes the task of quantitatively decoupling the activity of the particle's surface and interior, overcoming the cognitive bias of traditional detection methods that equate surface activity with overall activity, achieving the goal of accurately identifying the true reactive activity inside the particle, and eliminating the risk of misjudgment of activity caused by the asymmetry of core-shell structural information.

[0031] Current technologies generally rely on near-infrared spectroscopy to detect the activity of light-burned magnesia raw materials. This technology can only penetrate a shallow region of 50 to 100 micrometers on the particle surface, failing to reach the internal core of particles with a diameter of 0.5 to 2 millimeters. This physical limitation of surface detection means that the system can only acquire activity information of the overburned surface layer, completely masking the potentially high-activity characteristics of the underburned core region, resulting in a serious distortion of the overall activity assessment of the raw material. When particles with standard surface activity but hidden high activity enter the reaction system, the sudden exposure of the core in the later stages of the reaction can cause a sharp rise in temperature and pressure, leading to a violent boiling accident and the formation of anhydrous magnesium sulfate. Compared with the above limitations, the collaborative work of the terahertz tomography unit and the dual-zone activity decoupling unit constructs a completely new system for understanding the activity of raw materials. The tomography unit utilizes the strong penetrating characteristics of terahertz electromagnetic waves to achieve, for the first time, non-invasive acquisition of the dielectric response inside light-burned magnesia particles. Through tomographic reconstruction, it accurately locates the core-shell interface, expanding the detection dimension from the surface to the whole. Based on this, the dual-zone activity decoupling unit establishes a surface-to-shell activity mapping mechanism, quantifying the independent activity coefficients of the overburned surface layer and the underburned core region, and constructing a realistic core-shell activity decoupling model. This surface-to-shell detection method completely overcomes the blind spots of the core-shell structure, enabling the system to identify high-activity risks internally in advance, raising the accuracy of raw material activity assessment to a new level, laying a reliable data foundation for the stable control of the reaction process, and avoiding later-stage runaway and production accidents caused by misjudgment of activity.

[0032] Example 3: Please refer to Figure 1 The terahertz tomography unit includes a terahertz spectroscopy unit and a tomography reconstruction unit; The terahertz spectral unit is equipped with a transmission detection component and a pulse emission source. It emits terahertz electromagnetic waves that penetrate the entire scale of lightly calcined magnesium oxide particles, collects the time-domain electric field waveform after being modulated by different internal orientation regions, records the coordinated response characteristics of amplitude attenuation and phase delay, generates raw spectral data containing internal dielectric information, and transmits it to the tomographic reconstruction unit. The tomographic reconstruction unit is connected to the terahertz spectral unit to receive the raw spectral data. Based on the filtered back-projection tomographic reconstruction algorithm, it calculates the three-dimensional distribution of the dielectric constant inside the particle. It identifies and extracts the residual magnesium hydroxide or magnesium carbonate fingerprint lines in the core region through characteristic absorption peaks, locates the core-shell interface and internal heterogeneous structural features, and outputs structural parameters to the dual-zone active decoupling unit.

[0033] In this embodiment, the terahertz spectral unit is equipped with a transmission detector and a pulse emission source. It emits electromagnetic waves that penetrate the entire particle scale, acquiring the time-domain electric field waveform modulated by different internal orientation regions. The response characteristics of amplitude attenuation and phase delay are recorded, generating raw spectral data containing internal dielectric information. This unit completes the non-destructive acquisition of the particle's internal electromagnetic response signal, overcoming the penetration depth limitations of optical detection, and achieving the goal of obtaining the overall dielectric information of the particle, providing a raw data foundation for internal structure reconstruction.

[0034] The tomographic reconstruction unit is connected to the spectral unit, receiving raw spectral data and calculating the three-dimensional distribution of the dielectric constant inside the particle based on a filtered back-projection tomographic reconstruction algorithm. It then identifies and extracts residual magnesium hydroxide or magnesium carbonate fingerprint lines in the core region through characteristic absorption peaks, locating the core-shell interface and internal structural features. This unit completes the analysis and reconstruction of the raw spectral data, establishing a three-dimensional distribution map of the dielectric properties inside the particle. This achieves the goal of identifying the core-shell boundary and residual core precursors, providing structural localization evidence for active decoupling.

[0035] Existing near-infrared spectroscopy can only acquire shallow information from the particle surface, failing to reach the internal core region. This results in the overburned surface layer masking the high-activity characteristics of the underburned core region, leading to distorted raw material activity assessments. The terahertz spectroscopy unit utilizes the strong penetrating power of electromagnetic waves into non-metallic media, emitting ultrashort pulses to completely penetrate the entire particle, acquiring the modulated electric field waveform, and recording amplitude attenuation and phase delay, thus achieving complete capture of the internal dielectric response information. The tomographic reconstruction unit inverts the original spectrum based on a tomographic reconstruction algorithm, reconstructing the three-dimensional distribution of the internal dielectric constant, and identifying and locating the core-shell interface and internal heterogeneous structures through characteristic absorption peaks. The collaborative work of these two units constructs a complete technical chain from signal acquisition to structural reconstruction, extending the detection dimension from the surface to the entire particle, overcoming the blind spot in core-shell structure perception, and enabling the identification of the boundary positions between the surface and the core, as well as internal characteristics. This provides a true structural parameter basis for subsequent dual-zone activity decoupling, eliminating the risk of misjudgment of activity due to limitations in surface detection, and achieving the visualization and localization of the internal activity information of lightly burned magnesium oxide particles.

[0036] Example 4: Please refer to Figure 1 The dual-region active decoupling unit includes a feature fusion unit and an active inversion unit; The feature fusion unit is connected to the terahertz tomography unit and the surface detection unit. It receives the internal dielectric constant distribution and the absorption spectrum of surface functional groups, establishes a spatiotemporal alignment mapping of the core and shell features, constructs a fusion dataset related to the core and shell positions and surface activity, and outputs it downstream. The activity inversion unit is connected to the feature fusion unit. Based on the fusion dataset, a dielectric response-activity mapping between the overburned surface layer and the underburned core region is constructed. The surface activity coefficient and the core activity coefficient are inverted through the difference in dielectric constant. A core-shell activity decoupling model is established and dual activity parameters are output downstream.

[0037] In this embodiment, the feature fusion unit is connected to both the terahertz tomography unit and the upstream surface detection unit, simultaneously receiving the dielectric constant distribution inside the particle and the absorption spectra of surface functional groups. By establishing a spatiotemporal alignment mapping of core-shell features, this unit correlates and fuses the core-shell structure position with surface activity information, constructing a fused dataset containing the overall activity characteristics of the particle. This process completes the integration of internal structural information and surface chemical information, breaking the separation between core and shell data, achieving the goal of establishing a one-to-one correspondence between particle core and shell features, and providing a complete data input foundation for subsequent activity decoupling.

[0038] The activity inversion unit is connected to the feature fusion unit, and a dielectric response activity mapping model between the overburned surface layer and the underburned core region is constructed based on the fused dataset. By analyzing the difference in dielectric constant, the surface activity coefficient and the core activity coefficient are calculated, establishing a core-shell activity decoupling model and outputting two parameters downstream. This process achieves the quantitative differentiation of particle surface and internal activity, overcoming the cognitive bias of equating surface activity with overall activity in traditional detection methods. It achieves the goal of accurately identifying the true reactive activity within the particles, fundamentally eliminating the risk of misjudgment of activity caused by information asymmetry in the core-shell structure.

[0039] Compared to existing near-infrared surface detection methods, which are generally used, the limited physical penetration depth of these methods only allows for the acquisition of shallow information on the particle surface, failing to reach the internal core region. This results in the overburned surface layer masking the high-activity characteristics of the underburned core region, causing distortion in the overall activity assessment of the raw material. Consequently, this leads to a sudden exposure of the core in the later stages of the reaction, resulting in a drastic temperature rise and phase degradation. This parent unit, through the collaboration of a feature fusion unit and an activity inversion unit, constructs a complete technical chain from data acquisition to model output. The feature fusion unit establishes a connection between internal torsional structure information and surface spectrochemical information, creating a spatiotemporal mapping between core-shell position and surface activity, thus solving the problem of data separation between the surface and interior. Based on this, the activity inversion unit constructs a mapping relationship between dielectric response and activity, calculating the surface activity coefficient and core activity coefficient separately through the difference in dielectric constant, achieving quantitative decoupling of the activity between the particle surface and interior. This technical approach breaks through the limitations of traditional single-point detection, expanding the activity assessment dimension from a single surface to a dual-region core-shell system, outputting true dual-activity parameters, and overcoming the blind spot of the core-shell structure. The system is thus able to identify high internal activity risks in advance, avoiding later-stage runaway due to surface masking, providing precise data support for the stable control of the reaction process, and improving the accuracy of raw material activity assessment and production safety.

[0040] Example 5: Please refer to Figure 1 The transition early warning module includes a thermal acoustic monitoring unit and a transition prediction unit; The thermoacoustic monitoring unit is configured on the inner wall of the reactor and in the stirring shaft area. It integrates a highly sensitive micro-heat detection component and a broadband ultrasonic pulse transceiver component. In the initial stage of the reaction, it synchronously collects the weak heat flow changes and slurry sound velocity fluctuations during the dissolution process of the surface overburnt layer. It constructs a time-series dataset containing the heat of dissolution phase change and acoustic propagation characteristics and outputs it downstream. The transition prediction unit is connected to the thermoacoustic monitoring unit and the upstream terahertz perspective module, respectively. It receives time-series datasets and constructs a core pre-release index calculation model by combining the surface activity coefficient and the core activity coefficient. It analyzes the collapse dynamic trend of the surface overburnt layer, predicts the exposure time of the core high-activity zone and the equivalent activity transition amplitude, and outputs transition warning information to the dual-zone reconstruction module.

[0041] In this embodiment, a thermoacoustic monitoring unit is configured on the inner wall of the reactor and in the stirring shaft area to collect micro-thermal changes and slurry acoustic velocity fluctuations during the dissolution process of the surface overburnt layer in the initial stage of the reaction, constructing a time-series dataset and outputting it downstream. This unit simultaneously acquires the thermal effects and acoustic response characteristics of the dissolution process in the early stage of the reaction, overcoming the lag limitation of single temperature monitoring and providing early data support for predicting active transitions.

[0042] The transition prediction unit is connected to the thermoacoustic monitoring unit and the terahertz imaging module. It receives time-series datasets, surface activity coefficients, and core activity coefficients, constructs a core pre-release index calculation model, predicts the exposure time and activity transition amplitude of the core high-activity region, and outputs transition warning information to the dual-zone reconstruction module. This unit correlates the particle surface and interior activity parameters with real-time dissolution kinetics, achieving a leap from symptom identification to risk prediction, and reserving sufficient time window for adjusting control strategies.

[0043] Existing technologies rely on temperature and pressure threshold alarms, which suffer from delayed response when the core is suddenly exposed in the later stages of a core-shell structure reaction, leading to violent boiling and phase degradation. Thermoacoustic monitoring units collect weak thermoacoustic signs of surface layer dissolution at the initial stage of the reaction, acquiring early information that conventional sensors cannot detect. Transition prediction units fuse thermoacoustic data with surface and internal activity coefficients to create a model, quantitatively analyzing surface layer collapse trends and predicting core exposure risks in advance. The collaborative work of these two units transforms the response from delayed to proactive, enabling the control system to shift from passive to active pre-adjustment, effectively preventing sudden loss of control and quality accidents in the later stages of the reaction.

[0044] Example 6: Please refer to Figure 1 The thermoacoustic monitoring unit includes a heat detection unit and an acoustic wave detection unit; The heat detection unit is arranged in the area near the liquid surface on the inner wall of the reactor. It integrates a high-sensitivity micro-heat sensor array to continuously collect the weak heat flow changes and phase change heat effects during the dissolution process of the surface overburnt layer in the initial stage of the reaction. It extracts the thermal characteristic information of dissolution kinetics and outputs it synchronously with the acoustic wave detection unit. The acoustic detection unit is located near the impeller area of ​​the stirring shaft. It is equipped with a broadband ultrasonic pulse transceiver array, which transmits detection pulses to the reaction slurry and receives the propagation echo modulated by the particle interface. It monitors the sound velocity fluctuation and energy attenuation characteristics, extracts the acoustic response of surface layer collapse, and merges it with thermal characteristics for downstream output.

[0045] In this embodiment, the heat detection unit is arranged in the near-liquid surface region of the inner wall of the reactor. During the initial stage of the reaction, it continuously collects the heat flow changes and phase transition heat effects during the dissolution process of the surface overheated layer, extracting the thermal characteristics of the dissolution kinetics. This unit completes the real-time capture of the surface layer dissolution heat effect, overcoming the limitation of slow response in traditional temperature sensors, and achieving the goal of obtaining early thermal signs of dissolution, providing a thermal data foundation for subsequent fusion analysis.

[0046] The acoustic detection unit is positioned near the impeller on the stirring shaft. It emits probe pulses into the reaction slurry and receives the propagating echoes modulated by the particle interface, monitoring sound velocity fluctuations and energy attenuation characteristics to extract the acoustic response of surface layer collapse. This unit completes the acoustic characterization of surface layer structure collapse, identifies the dissolution state of the particle surface layer through sound velocity changes, achieves the goal of obtaining early acoustic signs of dissolution, and realizes data fusion by synchronously outputting thermal characteristics.

[0047] Existing technologies generally rely on threshold alarms based on traditional parameters such as temperature and pressure. When dealing with core-shell structured materials, they can only respond passively and cannot identify subtle signs of surface layer dissolution in the early stages of the reaction, leading to control lag when the core is suddenly exposed. A combined thermoacoustic monitoring architecture is constructed through the collaborative work of a thermal detection unit and an acoustic detection unit. The thermal detection unit continuously collects heat flow changes and phase transition heat effects during the dissolution process of the surface overburnt layer, overcoming the sensitivity limitations of single temperature monitoring and capturing the thermodynamic characteristics of the early dissolution stage. The acoustic detection unit transmits probe pulses and receives echoes modulated by the particle interface, monitoring sound velocity fluctuations and energy attenuation to identify the acoustic characteristics of surface layer collapse. The synchronous output of both units achieves the fusion of thermal effects and acoustic responses, constructing a time-series dataset containing dissolution phase transition heat and structural collapse characteristics. This provides the transition prediction unit with complete early warning information, enabling the system to identify the surface layer dissolution state at the initial stage of the reaction, overcoming the lag of traditional detection methods and achieving advanced perception of active transition risks.

[0048] Example 7: Please refer to Figure 1 The transition prediction unit includes an index construction unit and a trend prediction unit; The index construction unit is connected to the thermoacoustic monitoring unit and the dual-zone activity decoupling unit respectively. It receives the thermoacoustic fusion dataset and the surface-shell activity coefficients, establishes the correlation mapping between surface dissolution and core exposure based on the core-shell interface, constructs the core pre-release index model, and outputs it to the trend prediction unit. The trend prediction unit is connected to the index construction unit and the dual-zone reconstruction module respectively. Based on the core pre-release index, it quantitatively analyzes the dynamic trend of surface overburning layer dissolution and collapse, calculates the exposure time and equivalent activity transition amplitude of the core high-activity zone, generates transition early warning information containing transition time and amplitude, and transmits it to the dual-zone reconstruction module.

[0049] In this embodiment, the index construction unit is connected to both the thermoacoustic monitoring unit and the dual-zone activity decoupling unit, simultaneously receiving the thermoacoustic fusion dataset along with the surface activity coefficient and the core activity coefficient. This unit establishes a correlation mapping between the surface dissolution process and the core exposure risk based on the core-shell interface location, fusing real-time dissolution kinetics characteristics with particle surface and interior activity parameters to construct a core pre-release index calculation model. This process achieves a deep correlation between thermoacoustic data and activity coefficients, overcoming the limitations of single physical field monitoring, and achieving the goal of establishing a quantitative relationship between the dissolution process and internal activity release, providing a reliable computational framework for trend prediction.

[0050] The trend prediction unit is connected to both the index construction unit and the dual-zone reconstruction module, and quantitatively analyzes the dynamic trend of surface overburnt layer dissolution and collapse based on the core pre-release index. This unit analyzes the evolution of the surface layer thickness decay rate and the core exposure probability, calculates the specific moment when the highly active core region is about to be exposed and the equivalent active transition amplitude, generates early warning information containing the transition moment and amplitude, and transmits it to the dual-zone reconstruction module. This process completes the transition from symptom identification to risk quantification, changing the passive mode of traditional endpoint detection, achieving the goal of predicting active transition risks in advance, and reserving sufficient time window for adjusting control strategies.

[0051] Current technologies generally employ near-infrared surface detection combined with temperature and pressure threshold alarms, which have dual limitations when dealing with core-shell structured raw materials. On the one hand, surface detection cannot identify the highly active internal core, leaving the system unprepared when the core is suddenly exposed in the later stages of the reaction. On the other hand, traditional monitoring can only passively alarm after boiling occurs, failing to identify early signs of surface layer dissolution. The collaborative work of the index construction unit and the trend prediction unit reconstructs the technical path for risk prediction. The index construction unit integrates the thermoacoustic characteristics of surface layer dissolution obtained from thermoacoustic monitoring with the surface-to-internal activity coefficients output by terahertz radiography, establishing a correlation mapping between the surface dissolution process and the risk of core exposure, enabling the system to assess the probability of internal activity release based on real-time dissolution status. The trend prediction unit analyzes the dissolution and collapse trend of the overburned surface layer, quantitatively calculating the core exposure time and activity transition amplitude, achieving a shift from delayed response to proactive early warning. The synergy of these two methods allows the system to predict risks in the later stages of the reaction from the initial stage, breaking the passivity of traditional detection methods, effectively avoiding boiling runaway and phase degradation caused by sudden core exposure, and significantly improving the safety and controllability of the formation process.

[0052] Example 8: Please refer to Figure 1 The dual-zone reconstruction module includes a dual-zone modeling unit and a dynamic switching unit; The dual-zone modeling unit is connected to the terahertz perspective module and the transition early warning module respectively. It receives the surface activity coefficient, core activity coefficient and core pre-release index. Based on the heterogeneous divide-and-conquer strategy, it constructs a coupled framework of surface overburnt layer shrinkage and dissolution and core underburnt zone reaction kinetics. It analyzes the evolution law of interface mass transfer resistance and core-shell exposed area in real time, generates the dual-zone heterogeneous model skeleton under the current working condition and outputs it to the dynamic switching unit. The dynamic switching unit is connected to the dual-zone modeling unit and the closed-loop control module respectively. It receives the dual-zone heterogeneous model skeleton and core pre-release index, monitors the core pre-release index threshold determination result, triggers smooth switching of the control domain when it predicts that the core is about to be exposed, performs hot-plugging of model structure and online reloading of reaction rate constant and interface mass transfer coefficient, and outputs the dual-zone coupling control parameter set after switching downstream.

[0053] In this embodiment, the dual-zone modeling unit is connected to the terahertz perspective module and the transition early warning module, respectively. It receives the surface activity coefficient, core activity coefficient, and core pre-release index. Based on a heterogeneous divide-and-conquer strategy, it constructs the control equations for the surface overburnt layer shrinkage and dissolution and the reaction kinetic equations for the core underburnt zone, respectively. It analyzes the evolution of interfacial mass transfer resistance and core-shell exposed area in real time, generates the framework of the dual-zone heterogeneous model, and outputs it to the dynamic switching unit. This unit completes the modeling task of the heterogeneous reaction characteristics of the core-shell structure, overcoming the simplification limitations of traditional single-zone homogeneous models. It achieves the goal of realistically representing the independent reaction processes of the surface and interior dual zones, providing a precise structured model foundation for dynamic control.

[0054] The dynamic switching unit is connected to both the dual-zone modeling unit and the closed-loop control module. It receives the dual-zone heterogeneous model skeleton and core pre-release index, monitors the core pre-release index threshold determination result, and triggers a smooth switching of the control domain when the core is predicted to be exposed. This involves hot-plugging the model structure and online reloading of the reaction rate constant and interface mass transfer coefficient, outputting the switched dual-zone coupled control parameter set downstream. This unit completes the adaptive switching task from single-zone control to dual-zone control, changing the rigid mode of traditional fixed-parameter models. It achieves the goal of adjusting the control strategy in real time according to the core-shell exposure state, ensuring accurate matching between control parameters and the actual activity state of the material.

[0055] Existing technologies generally employ single-region homogeneous reaction models based on near-infrared surface detection. These models assume a uniform distribution of raw material activity and cannot characterize the heterogeneous reaction characteristics of the overburned surface layer and the underburned core region of core-shell structured raw materials. When the overburned surface layer slowly dissolves, the reaction rate calculated by the single-region model based on the surface activity coefficient is far lower than the actual reaction rate after core exposure, leading to a significant lag in the acid addition strategy and triggering violent temperature rises and boiling over when the core is suddenly exposed. The collaborative work of a dual-region modeling unit and a dynamic switching unit reconstructs the technical architecture of reaction control. The dual-region modeling unit establishes a core shrinkage model for the surface dissolution zone and kinetic equations for the core reaction zone based on a heterogeneous divide-and-conquer strategy. By analyzing the evolution of interfacial mass transfer resistance and core-shell exposed area in real time, it constructs a dual-region heterogeneous model framework that can realistically characterize the heterogeneous reaction of the core and shell, upgrading the digital twin model from a single-region homogeneous assumption to an independent evolution of the surface and core regions. Based on this, the dynamic switching unit monitors the core pre-release index and triggers a smooth switch of the electrochemical reaction control domain when the core is about to be exposed. This enables hot-swapping of the model structure and online parameter reloading, transforming the control strategy from a single-region model based on surface activity to a coupled model based on both surface and shell activity. The synergy of these two mechanisms allows the system to dynamically adjust the control equations and parameters according to the real-time state of core-shell exposure, completely solving the failure problem of single-region models in core-shell structured raw material applications and achieving precise control of the reaction process.

[0056] Example 9: Please refer to Figure 1 The dual-zone modeling unit includes a partitioned modeling unit and a coupled analytical unit; The partitioned modeling unit is connected to the terahertz perspective module, receives the surface activity coefficient and the core activity coefficient, constructs the control equation for the nucleus shrinkage and dissolution of the overburned surface layer and the first-order reaction kinetic equation for the underburned core region, independently analyzes the reaction rate characteristics and phase transition boundary conditions of the two regions, generates heterogeneous sub-models of the surface region and the core region and outputs them to the coupled analytical unit. The coupled analytical unit is connected to the partitioned modeling unit and the transition early warning module. It receives the heterogeneous sub-model and the core pre-release index, establishes the correlation between the interfacial mass transfer resistance of the surface dissolution zone and the core reaction zone, analyzes the migration trajectory of the core-shell interface and the evolution law of the core exposed area in real time, constructs a dual-zone dynamic coupling skeleton, and outputs the model structure to the dynamic switching unit.

[0057] In this embodiment, the region modeling unit receives the surface activity coefficient and core activity coefficient output by the terahertz perspective module. It constructs the core shrinkage dissolution control equation and reaction kinetic equation for the heterogeneous reaction mechanism of the overburned surface layer and the underburned core region, respectively, and independently analyzes the reaction rate characteristics and phase transition boundary conditions of the two regions. This unit completes the divide-and-conquer modeling task of the heterogeneous reaction characteristics of the core-shell structure, overcoming the simplistic limitation of traditional single-region homogeneous models that confuse surface and core activities. It achieves the goal of realistically representing the independent evolution of surface dissolution and core reactions, providing a heterogeneous sub-model foundation for subsequent coupled analysis.

[0058] The coupling analysis unit receives the heterogeneous sub-model output by the partitioned modeling unit and the core pre-release index transmitted by the transition early warning module. It establishes the correlation between the interfacial mass transfer resistance between the surface dissolution zone and the core reaction zone, and tracks the migration trajectory of the core-shell interface and the evolution of the core exposed area in real time. This unit completes the analysis of the dynamic coupling mechanism of the core-shell interface, overcomes the limitation that independent sub-models cannot characterize the interaction between the two zones, and achieves the goal of constructing a dynamic coupling framework for the surface and core zones, providing structured model support for adaptive switching of the control domain.

[0059] Compared to the single-region homogeneous reaction models commonly used in existing technologies, this parent unit achieves a fundamental upgrade in model architecture through the synergy of a partitioned modeling unit and a coupled analytical unit. Existing technologies, assuming uniform distribution of raw material activity, use a single kinetic equation to describe the overall reaction process, failing to identify the independent reaction characteristics of the overburned surface layer and the underburned core region. This leads to severe model prediction distortion when the core is suddenly exposed in the later stages of the core-shell structure raw material reaction. The partitioned modeling unit employs a heterogeneous divide-and-conquer strategy, establishing independent governing equations for surface core shrinkage and dissolution and the core reaction, realistically characterizing the heterogeneous reaction mechanism of the two regions. Building upon this, the coupled analytical unit introduces interfacial mass transfer resistance correlations, integrating the independent heterogeneous sub-models into a dynamic coupled framework through the core-shell interface migration and exposure area evolution laws, achieving a leap from static homogeneity to dynamic heterogeneity. The synergy of these two elements enables the digital twin model to simultaneously track the surface dissolution process and core exposure risk, realistically reflecting the independent evolution and interaction of the surface and core regions. This completely solves the failure problem of single-region models in core-shell structure raw material applications, providing a reliable model foundation for precise control and parameter adaptation of the reaction process.

[0060] Example 10: Please refer to Figure 1 The dynamic switching unit includes a threshold monitoring unit and a model switching unit; The threshold monitoring unit establishes a judgment logic based on the critical threshold of the core pre-release index. It continuously monitors the index change trajectory and growth rate through a sliding window algorithm. When it predicts that the core high-activity area is about to be exposed, it generates a switching trigger command to achieve a smooth transition prediction from the surface dissolution control domain to the dual-zone coupling control domain. The model switching unit receives the switching trigger command and the dual-zone heterogeneous model skeleton, performs hot-plugging of the digital twin model structure and smooth migration of the reaction control domain, reloads the reaction rate constant and interface mass transfer coefficient of the surface dissolution zone and the core reaction zone online, reconstructs the dual-zone coupled control parameter set and transmits it to the closed-loop control module.

[0061] In this embodiment, the threshold monitoring unit establishes critical threshold determination logic based on the core pre-release index, continuously monitors the index's change trajectory and growth rate, and generates a switching trigger command when it predicts that the core highly active region is about to be exposed. This unit completes the real-time determination of the critical state of reaction risk, overcoming the lag limitation of traditional temperature and pressure threshold alarms, achieving the goal of predicting the timing of control domain switching in advance, providing accurate timing signals for model switching, and realizing a smooth transition prediction from the surface dissolution control domain to the dual-zone coupled control domain.

[0062] The model switching unit receives the switching trigger command and the dual-zone heterogeneous model skeleton, performs hot-swapping of the digital twin model structure and smooth migration of the reaction control domain, and reloads the reaction rate constant and interfacial mass transfer coefficient of the surface dissolution zone and the core reaction zone online. This unit completes the real-time adaptive update of the control model and parameters, breaking through the rigidity limitation of traditional fixed-parameter models that cannot adapt to changes in the core-shell structure. It achieves the goal of dynamically adjusting the control strategy according to the core-shell exposure state, ensuring that the control parameters are accurately matched with the actual active state of the material, and transmits the reconstructed dual-zone coupled control parameter set to the closed-loop control module.

[0063] Compared to the single-zone homogeneous reaction model and fixed-parameter control strategy commonly used in existing technologies, this parent unit achieves a fundamental upgrade in control architecture through the collaboration of a threshold monitoring unit and a model switching unit. Existing technologies, due to their rigid model structures and fixed parameters, cannot adaptively adjust when the core is suddenly exposed in the later stages of a core-shell structure reaction, leading to severe lag in the control strategy and resulting in bumping and phase degradation. The threshold monitoring unit establishes a critical threshold judgment logic for the core pre-release index, continuously monitors the index change trajectory, and generates a switching trigger command before the core is about to be exposed, achieving a shift from passive alarm to proactive prediction. Based on this, the model switching unit performs hot-plugging of the model structure and smooth migration of the reaction control domain, online reloading of the reaction rate constant and interfacial mass transfer coefficient, enabling the control strategy to adapt from a single-zone mode based on surface activity to a coupled mode based on both surface and core activity. The collaboration of these two units allows the system to dynamically adjust the control equations and parameters according to the real-time state of core-shell exposure, completely solving the failure problem of fixed models in core-shell structure raw material applications, achieving precise adaptive control of the reaction process, and effectively avoiding bumping runaway and quality accidents caused by sudden core exposure.

[0064] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated units described above can be implemented in hardware or as software functional units. The above are merely embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made based on the description and drawings of this application, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0065] The specific embodiments of the invention have been described in detail above, but they are only examples, and this application is not limited to the specific embodiments described above. For those skilled in the art, any equivalent modifications or substitutions to the invention are also within the scope of this application. Therefore, all equivalent changes, modifications, and improvements made without departing from the spirit and principles of this application should be covered within the scope of this application.

Claims

1. A control system for the formation process of fertilizer-grade magnesium sulfate monohydrate based on digital twins, characterized in that: The system includes a raw material pre-screening module, a terahertz scanning module, a transition early warning module, a dual-zone reconstruction module, and a control verification module; The raw material pre-screening module is configured at the feed port of the light-burned magnesium oxide raw material silo. Based on the results of spectral detection, visual recognition and particle size analysis, it identifies and removes extreme particles with severe overburning or complete underburning, and sends medium core-shell structure particles into the terahertz perspective module. The terahertz perspective module is connected to the multimodal pre-screening module. It obtains dielectric response characteristics by penetrating the interior of the particles through terahertz electromagnetic waves, analyzes the surface activity coefficient and the core activity coefficient, and establishes a core-shell activity decoupling model. The transition early warning module is installed inside the reactor. During the initial stage of the reaction, it monitors the micro-thermal changes and ultrasonic pulse response during the dissolution of the overburned layer on the surface, generates the core pre-release index, and predicts the activity transition time by combining the surface activity coefficient and the core activity coefficient. The dual-zone reconstruction module is connected to the terahertz vision module and the early warning module respectively. It receives the surface activity coefficient, the core activity coefficient and the core pre-release index, and constructs a dual-zone coupled dynamic model of the surface dissolution zone and the core reaction zone. When the core is predicted to be exposed, the control equation and model parameters are dynamically switched. The control verification module is connected to the dual-zone reconstruction module. Based on the output of the dual-zone coupled dynamic model, it executes a segmented pre-adjustment control strategy and performs online verification of the phase state of the discharged crystal. The verification results are fed back to the terahertz imaging module to correct the detection parameters.

2. The fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 1, characterized in that, The terahertz fluoroscopy module includes a terahertz tomography unit and a dual-zone active decoupling unit; The terahertz tomography unit is located in the light-burned magnesium oxide particle transport channel. It emits terahertz electromagnetic waves to penetrate the particles and collect time-domain spectral signals. The internal dielectric constant distribution is analyzed by the tomographic reconstruction algorithm, and the characteristic absorption peaks of magnesium hydroxide or magnesium carbonate in the core region are extracted to determine the core-shell interface position and internal structural features. The dual-zone active decoupling unit receives the internal structural features output by the tomographic scanning unit and the surface functional group information output by the upstream surface detection unit, establishes an active mapping model between the overburned surface layer and the underburned core region, calculates the surface activity coefficient and the core activity coefficient, constructs a core-shell active decoupling model, and outputs dual parameters to the downstream module.

3. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twins according to claim 2, characterized in that, The terahertz tomography unit includes a terahertz spectroscopy unit and a tomography reconstruction unit; The terahertz spectral unit is equipped with a transmission detection component and a pulse emission source. It emits terahertz electromagnetic waves that penetrate the entire scale of lightly calcined magnesium oxide particles, collects the time-domain electric field waveform after being modulated by different internal orientation regions, records the coordinated response characteristics of amplitude attenuation and phase delay, generates raw spectral data containing internal dielectric information, and transmits it to the tomographic reconstruction unit. The tomographic reconstruction unit is connected to the terahertz spectral unit to receive the raw spectral data. Based on the filtered back-projection tomographic reconstruction algorithm, it calculates the three-dimensional distribution of the dielectric constant inside the particle. It identifies and extracts the residual magnesium hydroxide or magnesium carbonate fingerprint lines in the core region through characteristic absorption peaks, locates the core-shell interface and internal heterogeneous structural features, and outputs structural parameters to the dual-zone active decoupling unit.

4. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 3, characterized in that, The dual-region active decoupling unit includes a feature fusion unit and an active inversion unit; The feature fusion unit is connected to the terahertz tomography unit and the surface detection unit. It receives the internal dielectric constant distribution and the absorption spectrum of surface functional groups, establishes a spatiotemporal alignment mapping of the core and shell features, constructs a fusion dataset related to the core and shell positions and surface activity, and outputs it downstream. The activity inversion unit is connected to the feature fusion unit. Based on the fusion dataset, a dielectric response-activity mapping between the overburned surface layer and the underburned core region is constructed. The surface activity coefficient and the core activity coefficient are inverted through the difference in dielectric constant. A core-shell activity decoupling model is established and dual activity parameters are output downstream.

5. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 4, characterized in that, The transition early warning module includes a thermal acoustic monitoring unit and a transition prediction unit; The thermoacoustic monitoring unit is configured on the inner wall of the reactor and in the stirring shaft area. It integrates a highly sensitive micro-heat detection component and a broadband ultrasonic pulse transceiver component. In the initial stage of the reaction, it synchronously collects the weak heat flow changes and slurry sound velocity fluctuations during the dissolution process of the surface overburnt layer. It constructs a time-series dataset containing the heat of dissolution phase change and acoustic propagation characteristics and outputs it downstream. The transition prediction unit is connected to the thermoacoustic monitoring unit and the upstream terahertz perspective module, respectively. It receives time-series datasets and constructs a core pre-release index calculation model by combining the surface activity coefficient and the core activity coefficient. It analyzes the collapse dynamic trend of the surface overburnt layer, predicts the exposure time of the core high-activity zone and the equivalent activity transition amplitude, and outputs transition warning information to the dual-zone reconstruction module.

6. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 5, characterized in that, The thermoacoustic monitoring unit includes a heat detection unit and an acoustic wave detection unit; The heat detection unit is arranged in the area near the liquid surface on the inner wall of the reactor. It integrates a high-sensitivity micro-heat sensor array to continuously collect the weak heat flow changes and phase change heat effects during the dissolution process of the surface overburnt layer in the initial stage of the reaction. It extracts the thermal characteristic information of dissolution kinetics and outputs it synchronously with the acoustic wave detection unit. The acoustic detection unit is located near the impeller area of ​​the stirring shaft. It is equipped with a broadband ultrasonic pulse transceiver array, which transmits detection pulses to the reaction slurry and receives the propagation echo modulated by the particle interface. It monitors the sound velocity fluctuation and energy attenuation characteristics, extracts the acoustic response of surface layer collapse, and merges it with thermal characteristics for downstream output.

7. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 6, characterized in that, The transition prediction unit includes an index construction unit and a trend prediction unit; The index construction unit is connected to the thermoacoustic monitoring unit and the dual-zone activity decoupling unit respectively. It receives the thermoacoustic fusion dataset and the surface-shell activity coefficients, establishes the correlation mapping between surface dissolution and core exposure based on the core-shell interface, constructs the core pre-release index model, and outputs it to the trend prediction unit. The trend prediction unit is connected to the index construction unit and the dual-zone reconstruction module respectively. Based on the core pre-release index, it quantitatively analyzes the dynamic trend of surface overburning layer dissolution and collapse, calculates the exposure time and equivalent activity transition amplitude of the core high-activity zone, generates transition early warning information containing transition time and amplitude, and transmits it to the dual-zone reconstruction module.

8. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 7, characterized in that, The dual-zone reconstruction module includes a dual-zone modeling unit and a dynamic switching unit; The dual-zone modeling unit is connected to the terahertz perspective module and the transition early warning module respectively. It receives the surface activity coefficient, core activity coefficient and core pre-release index. Based on the heterogeneous divide-and-conquer strategy, it constructs a coupled framework of surface overburnt layer shrinkage and dissolution and core underburnt zone reaction kinetics. It analyzes the evolution law of interface mass transfer resistance and core-shell exposed area in real time, generates the dual-zone heterogeneous model skeleton under the current working condition and outputs it to the dynamic switching unit. The dynamic switching unit is connected to the dual-zone modeling unit and the closed-loop control module respectively. It receives the dual-zone heterogeneous model skeleton and core pre-release index, monitors the core pre-release index threshold determination result, triggers smooth switching of the control domain when it predicts that the core is about to be exposed, performs hot-plugging of model structure and online reloading of reaction rate constant and interface mass transfer coefficient, and outputs the dual-zone coupling control parameter set after switching downstream.

9. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 7, characterized in that, The dual-zone modeling unit includes a partitioned modeling unit and a coupled analytical unit; The partitioned modeling unit is connected to the terahertz perspective module, receives the surface activity coefficient and the core activity coefficient, constructs the control equation for the nucleus shrinkage and dissolution of the overburned surface layer and the first-order reaction kinetic equation for the underburned core region, independently analyzes the reaction rate characteristics and phase transition boundary conditions of the two regions, generates heterogeneous sub-models of the surface region and the core region and outputs them to the coupled analytical unit. The coupled analytical unit is connected to the partitioned modeling unit and the transition early warning module. It receives the heterogeneous sub-model and the core pre-release index, establishes the correlation between the interfacial mass transfer resistance of the surface dissolution zone and the core reaction zone, analyzes the migration trajectory of the core-shell interface and the evolution law of the core exposed area in real time, constructs a dual-zone dynamic coupling skeleton, and outputs the model structure to the dynamic switching unit.

10. A fertilizer-grade magnesium sulfate monohydrate formation process control system based on digital twin according to claim 7, characterized in that, The dynamic switching unit includes a threshold monitoring unit and a model switching unit; The threshold monitoring unit establishes a judgment logic based on the critical threshold of the core pre-release index. It continuously monitors the index change trajectory and growth rate through a sliding window algorithm. When it predicts that the core high-activity area is about to be exposed, it generates a switching trigger command to achieve a smooth transition prediction from the surface dissolution control domain to the dual-zone coupling control domain. The model switching unit receives the switching trigger command and the dual-zone heterogeneous model skeleton, performs hot-plugging of the digital twin model structure and smooth migration of the reaction control domain, reloads the reaction rate constant and interface mass transfer coefficient of the surface dissolution zone and the core reaction zone online, reconstructs the dual-zone coupled control parameter set and transmits it to the closed-loop control module.

Images