Intelligent production cooperative control system for slow release fertilizer

Abstract

The intelligent production cooperative control system of the slow-release fertilizer comprises a data acquisition module which acquires multi-source production data signals from batching, granulation, coating and drying sections in real time, and the multi-source production data signals at least comprise raw material ratio and material temperature; the intelligent optimization module receives the multi-source production data signal and releases a target based on a preset nutrient; the production execution control module receives the cooperative control instruction signal and generates a process control signal according to the cooperative control instruction signal; the digital twin module is respectively connected with the data acquisition module and the intelligent optimization module, receives the multi-source production data signal to update the state of the virtual production model, and provides an analog simulation and prediction feedback signal for the intelligent optimization module; the quality tracking module is connected with the data acquisition module and the production execution control module, and receives a multi-source production data signal and a process control signal. The slow release fertilizer intelligent production cooperative control system can solve the problems that slow release fertilizer production quality is not stable, energy consumption is high, and cooperative optimization cannot be achieved.

Description

Technical Field

[0001] This invention relates to the field of industrial automation and control technology, specifically to a collaborative control system for the intelligent production of slow-release fertilizers. Background Technology

[0002] Slow-release fertilizers are of great significance for improving fertilizer utilization, reducing environmental pollution, and achieving sustainable agricultural development. Their core value lies in the ability to match the nutrient release rate with the crop's needs. Currently, the production of slow-release fertilizers typically involves several key stages, including batching, granulation, coating, drying, and sieving. Among these, the coating process is crucial for controlling nutrient release characteristics. However, existing production processes generally face significant challenges: First, the production process is highly complex, with strong coupling relationships between multiple variables such as temperature, humidity, material viscosity, and coating amount. Relying on traditional discrete control systems or manual adjustments based on operator experience makes precise and stable control difficult, resulting in significant fluctuations and poor consistency in the quality of different batches of products, i.e., the nutrient release curves.

[0003] Secondly, most equipment on the production line, such as batching scales, granulators, coating machines, and dryers, operates independently, forming information silos and lacking a unified collaborative mechanism. This prevents coordinated optimization from the perspective of the entire process chain, leading not only to potential product quality issues but also to excessive energy and raw material consumption. Furthermore, product quality assessment is severely lagging, currently relying mainly on offline, post-production laboratory testing, which cannot provide real-time intervention and correction for quality deviations during production. Although some companies have attempted to introduce automation equipment in recent years, most have only achieved partial control of single points, failing to fundamentally solve the problems of whole-process collaborative optimization and real-time quality control. Therefore, the industry urgently needs a systematic solution that can connect the entire production line, achieve intelligent decision-making and collaborative control, and ensure stable and traceable product quality. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the purpose of this invention is to provide an intelligent collaborative control system for slow-release fertilizer production, which solves the problems of unstable production quality, high energy consumption, and inability to achieve collaborative optimization in slow-release fertilizer production. This invention addresses these problems by constructing a collaborative control system integrating data acquisition, intelligent optimization, execution control, virtual simulation, and quality tracking. The system collects production data from the entire process in real time. The intelligent optimization module performs global calculations based on the target release curve, generates collaborative instructions, and drives the linkage of equipment in each section. Simultaneously, the digital twin module performs pre-optimization in virtual space, and the quality tracking module provides closed-loop feedback, thereby achieving precise, stable, and efficient optimization of the production process.

[0005] This invention provides an intelligent collaborative control system for slow-release fertilizer production, comprising: The data acquisition module collects multi-source production data signals from the batching, granulation, coating and drying sections in real time. The multi-source production data signals include at least the raw material ratio, material temperature, ambient humidity, coating liquid flow rate and pressure. The intelligent optimization module receives multi-source production data signals and generates a set of collaborative control command signals based on the preset nutrient release target through the built-in collaborative optimization algorithm. The production execution control module receives collaborative control command signals and generates process control signals to drive the corresponding actuators in the batching, granulation, coating and drying sections. The digital twin module is connected to the data acquisition module and the intelligent optimization module respectively. It receives multi-source production data signals to update the state of the virtual production model and provides simulation and prediction feedback signals to the intelligent optimization module. The quality tracking module connects the data acquisition module and the production execution control module. It receives multi-source production data signals and process control signals, generates quality profile signals for product batch traceability and release curve prediction, and stores them persistently.

[0006] In one embodiment of the present invention, the data acquisition module further integrates an image acquisition unit, which is installed after the coating section to capture surface morphology image information of fertilizer particles to form a visual data signal. The multi-source production data signal further includes the visual data signal. The intelligent optimization module also receives the visual data signal and uses it as one of the key input factors for evaluating the uniformity and integrity of the coating. In turn, it fine-tunes and corrects the part of the cooperative control command signal related to the flow rate and pressure of the coating liquid to ensure the stability and consistency of the coating quality.

[0007] In one embodiment of the present invention, the collaborative optimization algorithm built into the intelligent optimization module is an advanced algorithm based on model predictive control. This algorithm predicts the behavior state of the system in the future by constructing a multivariate dynamic mathematical model of the production process, and takes the nutrient release target as the final optimization objective function. At the same time, it comprehensively considers multiple economic indicators such as production energy consumption and material consumption to perform multi-objective rolling optimization, thereby calculating a set of collaborative control command signals that can be globally optimal.

[0008] In one embodiment of the present invention, the production execution control module further includes an adaptive control unit, which can continuously receive real-time process parameter feedback signals from the data acquisition module and compare the feedback signals with the expected values ​​set by the collaborative control command signals in real time. Once the deviation is found to exceed the preset reasonable fluctuation range, the adaptive adjustment mechanism is immediately activated to generate a compensating process control signal to drive the actuator to move, thereby effectively suppressing external interference and internal parameter drift in the production process.

[0009] In one embodiment of the present invention, the virtual production model constructed by the digital twin module is a high-fidelity dynamic simulation model. This model continuously uses historical production data and real-time received multi-source production data signals to perform self-training and parameter updates through machine learning technology, enabling it to increasingly accurately simulate and map the real operating state of the physical entity, thereby providing the intelligent optimization module with more reliable and accurate simulation and prediction feedback signals.

[0010] In one embodiment of the present invention, the quality archive signal generated by the quality tracking module is a structured data set. This set not only includes the product batch number and the corresponding snapshot of all process control signals, but also integrates the key multi-source production data signals generated by the batch of products during the production process. Based on these data, the expected nutrient release pattern map of the batch of products is calculated through a pre-trained release curve prediction model, thereby realizing full-chain digital traceability from raw materials to final product quality prediction.

[0011] In one embodiment of the present invention, the system further includes a human-computer interaction and visualization module, which is connected to the intelligent optimization module and the quality tracking module respectively. It receives and visualizes collaborative control command signals, process control signals and quality file signals, and receives advanced production strategies and target adjustment commands input by operators and converts them into nutrient release target update signals that the intelligent optimization module can recognize, thereby realizing artificial intelligence-assisted decision-making and transparent management of the production process.

[0012] In one embodiment of the present invention, the data acquisition module further includes a data preprocessing and fusion unit, which is specifically responsible for cleaning, removing outliers, normalizing dimensions and aligning timestamps on the acquired raw multi-source production data signals. This unit fuses heterogeneous data from different sources, in different formats and with different acquisition frequencies into a standardized time-series data signal with consistency before transmitting it to the intelligent optimization module and the digital twin module, so as to ensure the high quality and high reliability of the data used for subsequent calculations and simulations.

[0013] In one embodiment of the present invention, when generating a collaborative control command signal, the intelligent optimization module will first call the simulation and prediction feedback signal provided by the digital twin module to perform simulation testing and effect prediction on multiple candidate collaborative control command combinations in virtual space, and select the candidate combination that can best approximate the nutrient release target and has the most stable process as the final collaborative control command signal, thereby realizing a simulation-driven risk-free pre-optimization control.

[0014] In one embodiment of the present invention, the connection between the production execution control module and the corresponding actuators in the batching, granulation, coating and drying sections adopts a hybrid network architecture composed of industrial Ethernet and real-time fieldbus. This architecture can ensure that process control signals are sent to various actuators with extremely low latency and high determinism, including servo drives for precisely controlling batching scales, electric actuators for adjusting valve openings, and frequency converters for controlling drum speeds, thereby realizing the coordinated linkage of the entire production system hardware. The present invention provides an intelligent collaborative control system for slow-release fertilizer production. This system addresses the aforementioned problems by constructing a collaborative control system integrating data acquisition, intelligent optimization, execution control, virtual simulation, and quality tracking. The system collects production data from the entire process in real time. The intelligent optimization module performs global calculations based on the target release curve, generating collaborative instructions to drive the linkage of equipment in each section. Simultaneously, the digital twin module performs pre-optimization in virtual space, and the quality tracking module provides closed-loop feedback, thereby achieving precise, stable, and efficient optimization of the production process. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a system architecture diagram for an intelligent collaborative control system for slow-release fertilizer production. Detailed Implementation

[0017] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0018] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Therefore, the drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0019] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0020] Please see Figure 1 The diagram illustrates the intelligent production collaborative control system for slow-release fertilizer of the present invention, comprising: a data acquisition module, which collects multi-source production data signals from the batching, granulation, coating, and drying stages in real time; the multi-source production data signals include at least raw material ratio, material temperature, ambient humidity, coating liquid flow rate, and pressure; an intelligent optimization module, which receives the multi-source production data signals and generates a set of collaborative control command signals based on a preset nutrient release target using a built-in collaborative optimization algorithm; a production execution control module, which receives the collaborative control command signals and generates process control signals to drive the corresponding actuators in the batching, granulation, coating, and drying stages; a digital twin module, which is connected to both the data acquisition module and the intelligent optimization module, receives multi-source production data signals to update the state of the virtual production model, and provides simulation and prediction feedback signals to the intelligent optimization module; and a quality tracking module, which is connected to both the data acquisition module and the production execution control module, receives multi-source production data signals and process control signals, generates quality profile signals for product batch traceability and release curve prediction, and stores them persistently.

[0021] Figure 1As shown, an unprecedented, highly integrated closed-loop control architecture has been constructed, completely breaking the "island" model of independent operation and information fragmentation in traditional fertilizer production. The system consists of five highly collaborative core modules, tightly connected through specific data signal flows, forming an organic whole capable of self-sensing, intelligent decision-making, precise execution, and continuous optimization. The data acquisition module, acting as the system's "sensory nerves," is responsible for capturing massive amounts of real-time operational data from the four most critical production stages: batching, granulation, coating, and drying. These multi-source production data signals are far from simple switching or analog quantities; they are a set of key process parameters that profoundly reflect the essence of the production state, including at least precise raw material ratio weight data, material temperature readings reflecting the progress of chemical reactions, environmental humidity values ​​affecting the coating curing effect, and instantaneous values ​​of coating liquid flow rate and pressure directly determining the uniformity of coating thickness. This raw data is transmitted in real-time and at high speed to the system's "intelligent brain"—the intelligent optimization module. This module is the decision-making center of the entire system. Its core value lies in the fact that it does not process individual parameters in isolation, but rather takes a holistic view based on a pre-set, highest-level instruction derived from agronomic needs—the nutrient release target (e.g., expecting fertilizer to release nutrients in an "S"-shaped curve over 120 days). Its built-in collaborative optimization algorithm is a complex mathematical calculation engine capable of simultaneously handling coupled variables from multiple processes. By solving multi-objective optimization problems, it generates a set of, rather than a single, collaborative control instruction signals. This set of instructions is a coordinated scheme that may simultaneously include adjusting the feeding speed of the batching scale, setting the temperature of the granulator, controlling the opening of the coating spray valve, and regulating the hot air volume of the dryer. Its purpose is to ensure that all processes move synchronously towards the same goal of achieving the final nutrient release characteristics. Next, this set of macro-level instructions is sent to the production execution control module, which acts as the "nerve endings" of the system. It translates the abstract instructions output by the optimization algorithm into directly executable process control signals that drive the action of specific physical actuators (such as servo motors, pneumatic control valves, and frequency converters), thereby completing the final closed loop from intelligent decision-making to change in the physical world.

[0022] Furthermore, to ensure the forward-looking nature and reliability of the optimization, the system introduces a digital twin module as a "pre-production sandbox." This module is bidirectionally connected to the data acquisition and intelligent optimization modules. It receives real-time data to drive its internal virtual production model, keeping it synchronized with the physical production line. This high-fidelity model can simulate the production results under different control strategies over a future period in virtual space, and informs the intelligent optimization module of the predicted outcomes in advance in the form of simulation and prediction feedback signals. This allows the module to conduct risk-free trial and error optimization, greatly improving the quality of decision-making. Finally, the quality tracking module, acting as the system's "black box" and quality auditor, simultaneously monitors the signal flow of the data acquisition module and the production execution control module. It links all "operation records" (process control signals) and "on-site conditions" (multi-source production data signals) during the production process of each batch of products, generating a structured quality archive signal that can be used for comprehensive traceability and quality prediction, and permanently storing it. These five modules are interconnected, forming a complete intelligent closed loop from perception to decision-making, execution, simulation, and traceability, achieving global collaborative optimization of the complex slow-release fertilizer production process.

[0023] In one embodiment of the present invention, the function of the data acquisition module is further deepened and expanded. The inventive point lies in adding advanced "machine vision" capabilities to the system, enabling it to directly examine the surface quality of products like the human eye. This claim specifies that the data acquisition module also integrates a dedicated image acquisition unit. This unit is not installed arbitrarily; its installation position is carefully designed and located at a key quality control point after the coating process and before the screening process. Its core task is to continuously capture high-definition surface morphology images of fertilizer granules that have just completed the coating process using a high-resolution industrial camera and a specially configured light source system, and to convert this image data into visual data signals that can be processed by algorithms. Thus, the connotation of the multi-source production data signal has undergone a qualitative leap; it is no longer merely composed of traditional physical parameters such as temperature, pressure, flow, and quantity, but has added visual data signals containing rich quality information.

[0024] like Figure 1As shown, these image data are synchronously transmitted to the intelligent optimization module, introducing a new and crucial dimension to the module's decision-making process. The image analysis algorithm built into the intelligent optimization module (usually based on a deep learning-based convolutional neural network model) processes these images in real time, quantitatively extracting key features related to coating quality, such as the integrity of the coating coverage, color uniformity, and the presence of defects such as cracks, drips, or adhesions. These quantitative results become key input factors for evaluating coating uniformity and integrity. The algorithm's decision logic thus becomes more complex and intelligent: it no longer relies solely on process parameters such as flow rate and pressure to indirectly infer quality, but can directly "see" the actual quality effect of the current output. Once visual analysis detects a trend of coating uniformity deviating from the set target, even if the flow rate and pressure parameters are still within the normal range, the intelligent optimization module will decisively fine-tune and correct the portion of the upcoming collaborative control command signal related to coating fluid flow rate and pressure. For example, if a thinner membrane is detected, the command will appropriately increase the flow rate; if a decrease in uniformity is detected, the atomization pressure or roller speed command may be adjusted. This creates a direct feedback loop based on the final product's quality and appearance, transforming quality control from offline, delayed manual sampling to online, real-time, and automatic precise regulation, greatly ensuring the stability and consistency of the coating quality of products produced at every moment.

[0025] like Figure 1As shown, focusing on the core of the intelligent optimization module—its built-in collaborative optimization algorithm—reveals the intrinsic mechanism by which the system achieves globally optimal decision-making. This claim defines the algorithm as an advanced algorithm based on model predictive control, marking a fundamental difference from traditional proportional-integral-derivative control or simple single-point feedback control. The core idea of ​​model predictive control lies in "multivariable, prediction, rolling optimization, and feedback correction." First, the algorithm's operation is based on a pre-established multivariable dynamic mathematical model of the production process. This model accurately describes complex coupling relationships and dynamic processes such as "how the film thickness will change in the next minute after the coating fluid flow rate increases, and what impact it will have on the drying section load." Within each control cycle, the algorithm uses this model to predict the system's behavior over a finite time period (prediction time domain) based on the current production state—that is, how key parameters will evolve if the current operation remains unchanged. Then, the algorithm compares this prediction result with the desired nutrient release target (which is transformed into a series of setpoint curves). It uses this objective as its fundamental optimization objective function, but its brilliance lies in not pursuing a single objective but also comprehensively considering multiple economic indicators such as production energy consumption and material consumption, performing multi-objective trade-offs and optimizations. For example, it might calculate an optimal path that satisfies the release curve requirements while minimizing steam consumption. Next, by solving a complex rolling time-domain optimization problem, it calculates a set of optimal coordinated control command signal sequences for the future control time domain (such as valve opening commands per second for the next 30 seconds), but only actually issues the first command in the sequence to the production line. In the next control cycle, the algorithm acquires the latest actual production data, refreshes the prediction model, and performs the above prediction-optimization process again, thus achieving "rolling optimization" based on the latest state. The significant advantage of this method is its ability to anticipate the future behavior of a multivariable coupled system, proactively taking control actions instead of passively waiting for deviations to occur before taking remedial measures. Furthermore, it can systematically handle the constraints and interactions between multiple work sections, thereby calculating the set of coordinated control command signals that achieves global optima rather than local optima, ultimately ensuring that production operates in a high-quality, low-consumption, and optimally stable state.

[0026] Furthermore, the digital twin module is the intelligent core and decision-making foundation of the system. The virtual production model constructed by this module is not a static, merely demonstrative model, but a high-fidelity, dynamic simulation model capable of real-time data interaction and bidirectional mapping with the physical production line. The model's excellence lies in its ability to continuously evolve and self-optimize, achieved through deep integration of machine learning technology. This technology allows the model to continuously train and update its parameters, like an increasingly experienced expert, utilizing a vast repository of historical production data and real-time multi-source production data signals received from the data acquisition module. Each production process provides new learning material for this virtual model. By analyzing the complex nonlinear relationships between different combinations of process parameters and actual output results, it continuously adjusts its internal algorithm weights and association rules, thereby continuously improving its simulation accuracy over time. This dynamic evolution capability ensures that the virtual model can increasingly accurately simulate and map the operating state, behavioral logic, and response characteristics of the physical production line in the real environment. Whether it's material flow, heat transfer, the progress of chemical reactions, or the formation of coatings, all can be highly reproduced in virtual space. It is precisely based on this high-precision mapping that the digital twin module can provide the intelligent optimization module with more reliable, accurate, and forward-looking simulation and predictive feedback signals. Before making decisions, the intelligent optimization module can first conduct countless trials and simulations in this "digital sandbox," rehearsing the execution consequences of various control strategies. This allows it to select the optimal solution before actual production, greatly reducing the trial-and-error costs and risks in actual production, and achieving a fundamental shift from experience-driven to model-driven and data-driven approaches. The core output of the quality tracking module is the aforementioned quality archive signal, which is essentially a highly structured, standardized, and information-rich data set. The design of this set goes far beyond a simple data listing; it constructs a complete digital product twin. It first contains the product's basic identity information, namely a unique batch number, like a product's ID card, binding all subsequent data to it. More importantly, it completely saves snapshots of all process control signals corresponding to this batch of products throughout the entire production process. This is equivalent to recording every "instruction" in producing this "product," serving as the original basis for reproducing the production process. In addition, the collection also deeply integrates all key multi-source production data signals generated during the production process of this batch of products. These data are "live recordings" of the production site, which truly reflect the actual effect of the equipment executing instructions and the instantaneous state of the production environment.Based on this massive amount of first-hand data, the quality tracking module invokes a pre-trained release curve prediction model. This model can comprehensively analyze numerous factors such as ingredient ratios, coating parameters, and temperature curves to calculate the expected nutrient release pattern of the batch of products over a future period, and store it in the archive as visualized data. This function is extremely significant because it means that even before the product has undergone lengthy laboratory testing, and even as it leaves the production line, the system can already make scientific and rapid predictions and assessments of its core performance characteristics—nutrient release characteristics. This achieves full-chain digital traceability, from raw material input to process parameters at each production stage, and finally to the prediction of final product quality. If quality issues or customer complaints arise later, this quality archive can be retrieved for precise retrospective analysis to quickly pinpoint the problem area. Simultaneously, this accumulation of high-quality data provides valuable data assets for continuous process improvement and product optimization.

[0027] like Figure 1 As shown, the human-computer interaction and visualization module is a crucial bridge connecting the intelligent decision-making system and human operational experts, greatly enhancing the system's usability and transparency. This module is not merely a passive information display, but a comprehensive platform integrating information presentation, status monitoring, interactive input, and decision support. By connecting to the intelligent optimization module and the quality tracking module respectively, it can receive and visualize key system signals in real time, such as collaborative control command signals calculated by the intelligent optimization module, process control signals issued by the production execution control module, and quality profile signals generated by the quality tracking module. Operators can clearly see the real-time operating status of the entire production line, the comparison between optimization targets and actual results, predicted product quality trends, and energy and material consumption through a centralized monitoring screen or remote terminal, thus gaining unprecedented overall insight into the production process and achieving transparent management. More importantly, this module has two-way interactive capabilities, allowing experienced operators or process engineers to input advanced production strategies and target adjustment instructions, such as temporarily changing the target release period based on order demand or fine-tuning optimization constraints based on raw material characteristics. This module converts these high-level instructions into structured nutrient release target update signals that the intelligent optimization module can accurately identify, and then transmits them to the optimization module as a new optimization benchmark. This perfectly combines human macro-level decision-making wisdom with the machine's micro-level precise control, forming a unique AI-assisted decision-making mode. Operators no longer need to tediously adjust isolated equipment parameters at the lower levels, but can directly focus on the performance goals of the final product. The system automatically decomposes these goals into specific executable actions, improving efficiency and reducing the demand on operators' low-level technical skills.

[0028] Specifically, the data preprocessing and fusion unit in the data acquisition module is the unsung hero behind the reliable operation of the entire system, acting as the "gatekeeper" to ensure data quality. In industrial environments, raw, multi-source production data signals directly acquired from sensors and equipment controllers are often coarse, noisy, incomplete, and even contradictory. This data may originate from different suppliers, follow different communication protocols, and have different sampling frequencies and units. Directly feeding it to upper-level intelligent algorithms for calculation is akin to "garbage in, garbage out," not only failing to yield accurate results but potentially leading to erroneous decisions. The core mission of this unit is to cultivate and manage this "raw jungle" of data. It first executes data cleaning and outlier removal algorithms, using statistical methods (such as the 3σ criterion) or model-based methods to identify and filter out or label jump signals, impulse interference, or constant values ​​caused by sensor failures that significantly exceed reasonable ranges. Next, it performs unit normalization processing on data from different sensors, converting all values ​​to the same scale to eliminate the impact of unit differences on subsequent multivariate analysis. Then, it addresses a crucial challenge—timestamp alignment. Because data acquisition times and transmission delays vary, this unit uses high-precision clock synchronization technology and data interpolation algorithms to unify all heterogeneous data streams onto the same timeline. This ensures that data analyzed at any given point in time was generated at the same or very close time, guaranteeing consistency in data timing. Ultimately, it fuses these carefully processed data from different sources, formats, and acquisition frequencies into a standardized time-series data signal with high consistency and reliability. Only after this rigorous process is this high-quality data transmitted to the intelligent optimization module and the digital twin module, providing a solid and reliable data foundation for all optimization calculations and simulation predictions, thus ensuring the scientific rigor and accuracy of the entire intelligent system's decisions from the outset.

[0029] The intelligent collaborative control system for slow-release fertilizer production of this invention addresses the aforementioned problems by constructing a collaborative control system integrating data acquisition, intelligent optimization, execution control, virtual simulation, and quality tracking. The system collects production data from the entire process in real time. The intelligent optimization module performs global calculations based on the target release curve, generating collaborative instructions to drive the linkage of equipment in each section. Simultaneously, the digital twin module performs pre-optimization in virtual space, and the quality tracking module provides closed-loop feedback, thereby achieving precise, stable, and efficient optimization of the production process.

[0030] Therefore, the intelligent production collaborative control system for slow-release fertilizer of the present invention can solve the problems of unstable production quality, high energy consumption, and inability to achieve collaborative optimization in slow-release fertilizer production.

[0031] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A controlled-release fertilizer intelligent production collaborative control system, characterized in that, include: The data acquisition module collects multi-source production data signals from the batching, granulation, coating and drying sections in real time. The multi-source production data signals include at least the raw material ratio, material temperature, ambient humidity, coating liquid flow rate and pressure. The intelligent optimization module receives the multi-source production data signal and generates a set of collaborative control command signals based on a preset nutrient release target through a built-in collaborative optimization algorithm. The production execution control module receives the collaborative control command signal and generates process control signals to drive the corresponding actuators in the batching, granulation, coating and drying sections. A digital twin module is connected to the data acquisition module and the intelligent optimization module respectively. It receives multi-source production data signals to update the state of the virtual production model and provides simulation and prediction feedback signals to the intelligent optimization module. The quality tracking module is connected to the data acquisition module and the production execution control module. It receives the multi-source production data signals and the process control signals, generates quality profile signals for product batch traceability and release curve prediction, and stores them persistently.

2. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The data acquisition module also integrates an image acquisition unit, which is installed after the coating section to capture surface morphology image information of fertilizer particles to form a visual data signal. The multi-source production data signal further includes the visual data signal. The intelligent optimization module also receives the visual data signal and uses it as one of the key input factors for evaluating the uniformity and integrity of the coating. It then fine-tunes and corrects the portion of the collaborative control command signal related to the flow rate and pressure of the coating liquid to ensure the stability and consistency of the coating quality.

3. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The intelligent optimization module incorporates a collaborative optimization algorithm, which is an advanced algorithm based on model predictive control. This algorithm predicts the system's behavior over a future period by constructing a multivariate dynamic mathematical model of the production process. It uses the nutrient release target as the final optimization objective function and comprehensively considers multiple economic indicators such as production energy consumption and material consumption to perform multi-objective rolling optimization, thereby calculating the globally optimal set of collaborative control command signals.

4. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The production execution control module also includes an adaptive control unit, which can continuously receive real-time process parameter feedback signals from the data acquisition module and compare the feedback signals with the expected values ​​set by the collaborative control command signals in real time. Once the deviation is found to exceed the preset reasonable fluctuation range, the adaptive adjustment mechanism is immediately activated to generate a compensating process control signal to drive the actuator to move, thereby effectively suppressing external interference and internal parameter drift in the production process.

5. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The virtual production model constructed by the digital twin module is a high-fidelity dynamic simulation model. This model continuously uses historical production data and real-time received multi-source production data signals to train itself and update its parameters through machine learning technology, enabling it to increasingly accurately simulate and map the real operating state of the physical entity, thereby providing the intelligent optimization module with more reliable and accurate simulation and prediction feedback signals.

6. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The quality tracking module generates a quality profile signal that is a structured data set. This set not only includes the product batch number and the corresponding snapshot of all process control signals, but also integrates the key multi-source production data signals generated during the production process of the batch of products. Based on this data, the expected nutrient release pattern map of the batch of products is calculated through a pre-trained release curve prediction model, thereby realizing full-chain digital traceability from raw materials to final product quality prediction.

7. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The system also includes a human-computer interaction and visualization module, which is connected to the intelligent optimization module and the quality tracking module respectively. It receives and visualizes the collaborative control command signals, process control signals and quality file signals. At the same time, it receives advanced production strategies and target adjustment commands input by operators and converts them into nutrient release target update signals that the intelligent optimization module can recognize, thereby realizing artificial intelligence-assisted decision-making and transparent management of the production process.

8. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The data acquisition module also includes a data preprocessing and fusion unit, which is specifically responsible for cleaning, removing outliers, normalizing dimensions, and aligning timestamps on the acquired raw multi-source production data signals. This unit fuses heterogeneous data from different sources, in different formats, and with different acquisition frequencies into a standardized time-series data signal with consistency before transmitting it to the intelligent optimization module and the digital twin module, ensuring the high quality and reliability of the data used for subsequent calculations and simulations.

9. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, When generating the collaborative control command signal, the intelligent optimization module will prioritize calling the simulation and prediction feedback signals provided by the digital twin module to conduct simulation tests and effect predictions on multiple candidate combinations of collaborative control commands in virtual space, and select the candidate combination that can best approximate the nutrient release target and has the most stable process as the final collaborative control command signal, thereby realizing a simulation-driven risk-free pre-optimization control.

10. The intelligent collaborative control system for slow-release fertilizer production according to claim 1, characterized in that, The connection between the production execution control module and the corresponding actuators in the batching, granulation, coating and drying sections adopts a hybrid network architecture composed of industrial Ethernet and real-time fieldbus. This architecture can ensure that the process control signals are sent to various actuators with extremely low latency and high determinism, including servo drives that precisely control the batching scale, electric actuators that adjust the valve opening, and frequency converters that control the drum speed, thereby realizing the coordinated linkage of the entire production system hardware.

Images