Non-woven fabric full-process remote intelligent control method and system based on Internet of Things

Abstract

The invention discloses a non-woven fabric whole-process remote intelligent control method and system based on the Internet of Things, and relates to the technical field of remote control, and the method comprises the steps: transmitting real-time process data to an edge server locally associated with a production line; the edge server matches and compares the received real-time process data with a pre-stored reference process digital twinborn model, and when a matching and comparison result indicates that at least one piece of process data deviates from a preset threshold value in the corresponding reference process digital twinborn model, the edge server sends the received real-time process data to the corresponding reference process digital twinborn model; the edge server calls a digital twinborn model associated with the deviation data to perform online simulation deduction so as to judge the process state deviation and obtain optimization parameters for correcting the deviation; and the edge server generates a control instruction based on the optimization parameters, and sends the control instruction to a corresponding execution mechanism on the production line through the Internet of Things so as to adjust spinning, cooling and web forming process links. The method has the effect of improving the remote control efficiency.

Description

Technical Field

[0001] This invention relates to the field of remote control technology, and in particular to a remote intelligent control method and system for the entire process of nonwoven fabric manufacturing based on the Internet of Things. Background Technology

[0002] In the production of nonwoven fabrics, especially high-end spunbond and meltblown materials, product quality is highly dependent on the precise and stable control of multiple core processes such as spinning, cooling, and web formation. Traditional production methods generally rely on mature equipment hardware and the on-site experience of operators, with their core technology lying in the continuous improvement of physical equipment and the static, experience-based setting of process parameters. However, as market demand evolves towards small-batch, multi-variety, and distributed flexible manufacturing, especially in emerging scenarios such as emergency material supply and regional customized supply, this traditional technology model based on centralized and experience-dependent methods is gradually revealing its inherent limitations.

[0003] Existing technologies typically limit process optimization to the mechanical design and offline simulation stages of the equipment itself. For example, for the coat hanger-type die flow channel that determines spinning uniformity, its design optimization largely relies on computer simulation and prototype testing during the R&D phase. Once put into use, its flow characteristics are fixed in the hardware structure and cannot be dynamically adjusted according to real-time raw material fluctuations or changes in the production environment. In the cooling and web-forming stages, although the duct design can be pre-optimized using fluid simulation, key states such as airflow pulsation and pressure distribution under the web during production can only be roughly monitored through periodic manual inspections or simple single-point sensors, lacking real-time, refined perception of the full-width, multi-variable flow field. This makes the production process control essentially an open-loop or weakly closed-loop state, with process adjustments lagging significantly behind the occurrence of quality deviations, making it difficult to meet the stringent requirements of process stability for high-speed, low-weight products. Furthermore, the operation and maintenance of the production line heavily relies on on-site diagnosis and debugging by experienced technicians. When production lines are distributed across diverse scenarios in different geographical locations, professional personnel cannot cover them in a timely manner, resulting in slow fault response, difficulty in standardizing and rapidly replicating process knowledge, and restricting the overall efficiency and reliability of the production network.

[0004] To address the aforementioned issues, there is an urgent need in this field for a technical solution that can deeply integrate process mechanisms with real-time production data to achieve remote sensing, intelligent analysis, and precise control throughout the entire process. This would break through the excessive reliance on fixed hardware and on-site manpower in the traditional model, thereby meeting the urgent needs of the nonwoven fabric industry to transform and upgrade towards flexible, intelligent, and networked manufacturing. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a remote intelligent control method and system for the entire process of nonwoven fabric manufacturing based on the Internet of Things.

[0006] In a first aspect, the present invention provides a remote intelligent control method for the entire process of nonwoven fabric manufacturing based on the Internet of Things, comprising the following steps: Real-time process data is acquired from multiple sensors deployed on the nonwoven fabric production line; the real-time process data includes spinning data characterizing the melt flow state, side blowing data characterizing the cooling airflow state, and web forming data characterizing the fiber web forming state. The real-time process data is sent to an edge server locally associated with the production line. The edge server matches and compares the received real-time process data with the pre-stored benchmark process digital twin model. The benchmark process digital twin model includes a coat hanger flow channel model based on non-Newtonian fluid simulation, a side-blowing flow field model based on computational fluid dynamics simulation, and a net-mounted suction wall flow collaborative model based on fluid numerical simulation. When the matching comparison result indicates that at least one process data deviates from the preset threshold in the corresponding benchmark process digital twin model, the edge server calls the digital twin model associated with the deviated data to perform online simulation and deduction, so as to determine the process state deviation and obtain the optimization parameters for correcting the deviation. The edge server generates control commands based on the optimized parameters and sends the control commands to the corresponding actuators on the production line via the Internet of Things to adjust the spinning, cooling and web forming processes.

[0007] Preferably, the acquisition of real-time process data from multiple sensors deployed on the nonwoven fabric production line includes: The spinning data is obtained from a sensor embedded in a coat hanger-type spinning die, and the spinning data includes at least melt temperature and melt pressure. The side-blowing data is acquired from a sensor array located on the air outlet surface of the side-blowing box, and the side-blowing data includes at least wind speed distribution and pressure distribution; The web formation data is acquired from a sensor located in the suction duct below the mesh curtain. The web formation data includes at least negative pressure gradient data and fiber web trajectory image data acquired from an image acquisition device facing the web formation area.

[0008] Preferably, the reference process digital twin model is pre-stored in the following manner: Non-Newtonian fluid simulation optimization of the coat hanger type spinning die flow channel was carried out by parametric modeling and genetic algorithm. The coat hanger type flow channel model was established, and the coat hanger type flow channel model defined the mapping relationship between the flow channel structure parameters and the melt flow uniformity. The side-blowing airflow field model is established by simulating the jet inside the side-blowing air box. The side-blowing airflow field model defines the correspondence between the flow guiding structure and the stable flow field pattern. By numerically simulating the interaction between airflow and fiber in the under-net suction system, a coordinated model of under-net suction and wall flow is established. This model defines the control relationship between suction pressure, wall-attached airflow, and fiber web forming uniformity.

[0009] Preferably, online simulation is performed by calling the digital twin model associated with the deviation data, including: When the spinning data deviates, the edge server calls the coat hanger-type flow channel model, and performs rheological calculations based on the real-time melt temperature and melt pressure, and dynamically outputs compensation parameters for the die head temperature zone as the optimization parameters.

[0010] Preferably, the online simulation and deduction by calling the digital twin model associated with the deviation data further includes: When the side-blowing data deviates, the edge server calls the side-blowing flow field model, performs pattern matching between the wind speed distribution and pressure distribution and the pre-stored stable flow field spectrum, and performs rapid simulation of the identified turbulent anomaly region based on the CFD reduced-order model, outputting the adjustment parameters for the fan frequency as the optimization parameters.

[0011] Preferably, the online simulation and deduction by calling the digital twin model associated with the deviation data further includes: When the negative pressure gradient data or fiber web trajectory image data in the web formation data indicate forming disturbance, the edge server calls the under-web suction wall flow collaborative model, integrates the negative pressure gradient data and fiber trajectory image data to perform gas-solid coupling simulation, and outputs the collaborative adjustment parameters for the main and auxiliary suction air volume and the opening of the attached airflow valve as the optimization parameters.

[0012] Preferred options also include: The cloud-based remote intelligent control platform receives and aggregates process data, alarm information, and optimization logs from multiple edge servers. The cloud platform maintains an adaptive process formula library based on the received data. The formula library contains a set of preset process parameters for different production scenarios. In response to a scenario selection instruction for a specific production line, the cloud platform retrieves the corresponding set of process parameters from the formula library and sends it to the edge server associated with the target production line to update the corresponding baseline process digital twin model.

[0013] Preferred options also include: The Internet of Things (IoT) continuously collects operational status data of key mechanical components in the production line; Based on the operational status data and historical fault records, the edge server uses a trend analysis algorithm to predict potential component failures and generate predictive maintenance early warning information. The predictive maintenance early warning information is sent to the designated maintenance management terminal.

[0014] Secondly, this invention provides a remote intelligent control system for the entire process of nonwoven fabric manufacturing based on the Internet of Things, including: The equipment layer includes spinning devices, side blowing devices, web forming and suction devices, multiple sensors integrated therein, and controlled actuators; The edge control layer includes an edge server and an IoT gateway deployed locally on the production line. The edge server is used to perform data reception, comparison, simulation and deduction and instruction generation. The IoT gateway is used to perform data communication between the device layer and the edge control layer. The cloud platform layer includes a remote intelligent control platform for cross-production line data aggregation, visualization, formula library management, and remote collaboration support. The edge control layer is connected to the device layer and the cloud platform layer via the Internet of Things, forming a closed loop of data interaction and control.

[0015] Thirdly, the present invention provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the IoT-based remote intelligent control method for the entire process of nonwoven fabrics as described above.

[0016] In summary, the present invention has at least one of the following beneficial technical effects: 1. This invention provides a remote intelligent control method for the entire process of nonwoven fabric manufacturing based on the Internet of Things. By deploying a sensor network across all dimensions on the production line and combining it with a digital twin model constructed based on precise physical mechanisms, it can perceive the microscopic state changes of core processes such as spinning, cooling, and web formation in real time. When the process parameters are detected to deviate from the threshold defined by the digital twin model, the corresponding model can be automatically and quickly invoked for online simulation and deduction, dynamically calculating the optimal correction parameters and issuing them for execution. This real-time closed-loop control of perception-analysis-decision-execution greatly overcomes the lag of traditional methods that rely on manual experience and static parameter settings, and significantly improves the accuracy, timeliness, and stability of process control. 2. The coat hanger-type flow channel model, side-blowing flow field model, and under-net suction wall flow collaborative model introduced in this invention can perform dynamic simulation and inversion based on real-time collected information such as melt state, airflow distribution, and fiber netting during the production process. This enables the production line to have an intelligent brain, adaptively respond to external interference and internal parameter drift, and achieve flexible expansion and dynamic optimization of the process window. 3. By combining IoT technology with edge computing, the edge server is responsible for real-time processing of local data, model matching, and rapid control, ensuring low latency and high reliability of response. The cloud platform aggregates data from multiple production lines, builds and maintains a shareable and iterative adaptive process formula library, so that even if the production lines are distributed in different regions, the optimal process formula can be remotely distributed through the cloud, and the baseline model and control logic on the edge side can be updated with one click. This not only greatly reduces the dependence on scarce on-site expert experience, but also realizes the rapid replication and inheritance of process knowledge. 4. By continuously collecting operational status data of key mechanical components through the Internet of Things, and based on historical data and trend analysis algorithms, it is possible to predict the performance degradation trend of components in the early stages before functional failures occur, and generate predictive maintenance warnings. This effectively avoids production losses and quality risks caused by unplanned downtime, extends equipment lifespan, optimizes spare parts inventory management, thereby comprehensively reducing production and maintenance costs and improving overall equipment efficiency and the overall reliability of the production network. Attached Figure Description

[0017] 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.

[0018] Figure 1 This is a flowchart of a method for remote intelligent control of the entire process of nonwoven fabric based on the Internet of Things, according to an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of a system for remote intelligent control of the entire process of nonwoven fabric based on the Internet of Things, according to an embodiment of the present invention. Detailed Implementation

[0020] The following is in conjunction with the appendix Figure 1-2 This application will be described in further detail.

[0021] In existing technologies, process control in nonwoven spunbond production lines largely relies on operators' on-site experience and independent adjustments of discrete parameters, making it difficult to achieve full-process collaborative optimization. Traditional methods, when faced with raw material fluctuations, environmental changes, or high-speed production, lack real-time insight and quantitative analysis of the coupling relationships between multiple physical fields such as melt rheology, airflow distribution, and fiber dynamics, resulting in delayed adjustments and insufficient accuracy. While existing monitoring systems can collect some data, they are typically only used for status display and over-limit alarms, unable to combine real-time data with in-depth process mechanism models for root cause diagnosis and proactive control. Especially when pursuing high-end products such as ultra-fine denier and low basis weight, the process window is narrow, making it difficult for traditional methods to stably maintain optimal production conditions, thus hindering further improvements in product quality and production efficiency.

[0022] To address the aforementioned issues, the inventors discovered that quality deviations in nonwoven fabric production are essentially the result of the coupling of multiple process stages deviating from their physical optimal points. During the research, it was found that the non-Newtonian flow characteristics of the melt within the coat hanger-type die, the complex turbulent structure in the side-blowing box, and the gas-solid two-phase flow interaction in the web-forming zone can all be accurately described using pre-established high-fidelity digital models. Therefore, a method was proposed to perform real-time matching and comparison between real-time sensor data and these benchmark process digital twin models. When a deviation is detected, instead of empirical trial-and-error adjustments, the corresponding digital model is triggered for online rapid simulation and deduction, reverse-engineering the root cause of the deviation and calculating precise compensation parameters. Further experimental and simulation verification led to the construction of a complete control logic from multi-source data perception and intelligent model diagnosis to closed-loop command execution, forming an adaptive optimization system driven by both data and models.

[0023] Specifically, the control system first uses sensor arrays deployed at key nodes of the production line to synchronously and in real-time collect multimodal process data, including temperature and pressure characterizing the melt state, flow field distribution characterizing cooling uniformity, and negative pressure gradient and fiber images characterizing web formation quality. After receiving the data, the edge server immediately performs high-speed matching and comparison analysis with pre-stored digital twin models of the coat hanger-type flow channel, the side-blowing flow field, and the under-web suction coordination. When the comparison results identify that data in a certain stage deviates from the ideal state defined by the model and exceeds a preset threshold, the system automatically calls the corresponding lightweight simulation engine. For example, it quickly solves the rheological equation based on real-time boundary conditions to inversely deduce the die head adjustment amount, or runs a reduced-order computational fluid dynamics model to simulate the air field adjustment effect, thereby obtaining quantitative optimized process parameters. Finally, these parameters are converted into control commands and sent to actuators such as the die head temperature control system, variable frequency fan, and airflow valves via the Internet of Things, achieving coordinated adjustment of the spinning, cooling, and web formation stages, forming a dynamic optimization closed loop.

[0024] Compared to existing technologies, traditional control methods rely on isolated parameter settings and manual intervention, making it difficult to address dynamic coupling issues under complex operating conditions. This solution innovatively constructs a real-time interactive and optimization closed loop between a "physical production line and a digital twin." By accumulating in-depth process knowledge into a computable model and utilizing real-time data for driving and calibration, it achieves a shift from passive monitoring to proactive optimization. Unlike existing discrete control systems, this solution possesses cross-process collaborative analysis and model simulation-based root cause diagnosis capabilities. It can dynamically generate optimal control strategies based on real-time operating conditions and ensure real-time response and global decision-making through an edge-cloud collaborative architecture. Through the above technical solutions, this application effectively overcomes the challenges of controlling multi-variable, strongly coupled, and fast-response processes in high-end nonwoven fabric production. While improving product uniformity and stability, it provides reliable core technical support for remote intelligent operation and maintenance and precise control in distributed and flexible production scenarios.

[0025] Example 1 This invention discloses a remote intelligent control method for the entire process of nonwoven fabric manufacturing based on the Internet of Things.

[0026] Reference Figure 1 A remote intelligent control method for the entire process of nonwoven fabric manufacturing based on the Internet of Things includes the following steps: Real-time process data is acquired from multiple sensors deployed on the nonwoven fabric production line; the real-time process data includes spinning data characterizing the melt flow state, side blowing data characterizing the cooling airflow state, and web forming data characterizing the fiber web forming state. The real-time process data is sent to an edge server locally associated with the production line. The edge server matches and compares the received real-time process data with the pre-stored benchmark process digital twin model. The benchmark process digital twin model includes a coat hanger flow channel model based on non-Newtonian fluid simulation, a side-blowing flow field model based on computational fluid dynamics simulation, and a net-mounted suction wall flow collaborative model based on fluid numerical simulation. When the matching comparison result indicates that at least one process data deviates from the preset threshold in the corresponding benchmark process digital twin model, the edge server calls the digital twin model associated with the deviated data to perform online simulation and deduction, so as to determine the process state deviation and obtain the optimization parameters for correcting the deviation. The edge server generates control commands based on the optimized parameters and sends the control commands to the corresponding actuators on the production line via the Internet of Things to adjust the spinning, cooling and web forming processes.

[0027] Specifically, this application further proposes a remote intelligent control method for the entire process of nonwoven fabrics based on the Internet of Things. By constructing a closed-loop control link of "perception-modeling-deduction-execution", it realizes the coordinated optimization and stable control of core process links such as spinning, cooling, and web formation.

[0028] Acquiring real-time process data from multiple sensors deployed on the nonwoven fabric production line refers to comprehensive and high-frequency digital sensing of key physical fields in the production process. Specifically, this involves simultaneously collecting multi-source heterogeneous data characterizing the melt rheological state, cooling airflow distribution, and fiber web formation morphology through temperature and pressure sensors deployed inside the coat hanger-type spinning die, array-type wind speed and pressure sensors on the air outlet surface of the side-blowing box, negative pressure sensors in the under-web suction duct, and visual sensors in the web-forming area. To ensure spatiotemporal consistency for subsequent joint analysis and model matching, the acquisition process of each sensor is synchronously triggered by a production line-level clock signal, and the raw data is uploaded in real time through an independent industrial bus or dedicated data channel.

[0029] Sending the real-time process data to an edge server locally associated with the production line is a crucial step in converging physical signals into digital information. The edge server, acting as the local computing hub of the production line, receives data streams from various sensors via a high-speed industrial network interface and performs preliminary verification, caching, and formatting, providing unified and standardized data input for subsequent intelligent analysis.

[0030] The edge server matches and compares the received real-time process data with a pre-stored benchmark process digital twin model, which is the core link in the entire solution to achieve intelligent judgment. The benchmark process digital twin model is a virtual standard built in the early stage through in-depth process simulation and optimization. For example, the coat hanger flow channel model based on non-Newtonian fluid simulation accurately describes the uniform distribution of ideal melt flow; the side-blowing flow field model based on computational fluid dynamics simulation defines a stable, pulsation-free ideal cooling airflow pattern; and the under-net adsorption wall flow collaborative model based on fluid numerical simulation describes the ideal web formation dynamics of fibers under the guidance of high negative pressure gradient and wall-attached airflow. The matching and comparison process is essentially a step-by-step, quantitative comparison and analysis of the real-time perceived "physical state" with the virtual "ideal standard".

[0031] When the matching comparison result indicates that at least one process data point deviates from a preset threshold in the corresponding baseline process digital twin model, the intelligent diagnosis and optimization process is triggered. At this time, the edge server calls the digital twin model associated with the deviation data to perform online simulation. This simulation is not a complete and time-consuming recalculation of the physical field, but a simplified solution or inverse calculation based on a pre-stored model. For example, for uneven spinning pressure, the system quickly inverts the equivalent resistance change or temperature field deviation in the flow channel based on real-time temperature and pressure data, thereby locating the root cause of the problem; for airflow pulsation, it calls a reduced-order computational fluid dynamics model to quickly simulate the impact of guide vane adjustment on turbulence in a specific area. Through this simulation, the system can go beyond simple threshold alarms and obtain optimized parameters for quantitatively correcting the process state deviation.

[0032] Finally, the edge server generates specific control commands based on the optimized parameters and sends these commands to the corresponding actuators on the production line via the Internet of Things (IoT). These commands may include fine-tuning the heating power of a specific temperature zone in the die head, setting the frequency of the inverter for the opposite blower, or coordinating the opening of the suction valve and the wall-mounted airflow valve. The execution of these commands directly affects the spinning, cooling, and web-forming processes, thus forming a complete closed loop from data perception and intelligent analysis to precise control.

[0033] Specifically, through the aforementioned collaborative technical solution, this application constructs an intelligent system capable of real-time insight and proactive optimization of complex processes. Synchronous acquisition and high-speed transmission from multiple sensors ensure the system's real-time and accurate perception of the entire production status. A benchmark digital twin model provides precise and quantitative evaluation criteria for process quality. Online simulation based on the model enables the system to possess root cause diagnosis and parameter optimization capabilities similar to expert experience. Finally, remote command issuance and execution via the Internet of Things (IoT) feeds optimization decisions back to the physical production line without delay.

[0034] Compared to traditional control methods that rely on human experience and make delayed adjustments, this solution deeply integrates data-driven and model-driven approaches, transforming post-event remediation into pre-event early warning and real-time in-process control. This effectively solves the challenge of achieving refined and consistent control of highly coupled multi-variable processes in high-speed, wide-width nonwoven fabric production. Through this closed loop, the system can automatically maintain the process state dynamically stable near the optimal baseline, significantly improving the production success rate and quality uniformity of high-end products, and providing a reliable technical foundation for remote monitoring and flexible operation of the production line.

[0035] Furthermore, the acquisition of real-time process data from multiple sensors deployed on the nonwoven fabric production line includes: The spinning data is obtained from a sensor embedded in a coat hanger-type spinning die, and the spinning data includes at least melt temperature and melt pressure. The side-blowing data is acquired from a sensor array located on the air outlet surface of the side-blowing box, and the side-blowing data includes at least wind speed distribution and pressure distribution; The web formation data is acquired from a sensor located in the suction duct below the mesh curtain. The web formation data includes at least negative pressure gradient data and fiber web trajectory image data acquired from an image acquisition device facing the web formation area.

[0036] Specifically, the above technical solution aims to clarify the precise locations and types of data acquisition, ensuring that the acquired data effectively characterizes the core state of the corresponding process stage. Specifically, the spinning data directly points to the interior of the melt flow channel, which determines fiber forming quality; temperature and pressure are key physical quantities reflecting the uniformity and stability of melt flow. Side-blowing data is acquired through a sensor array, aiming to capture the airflow field distribution across the entire width; wind speed and pressure distribution are direct indicators for evaluating cooling uniformity. Web-forming data combines the negative pressure gradient measured by physical sensors with the fiber trajectory acquired through image visualization. The former reflects the adsorption capacity distribution of the airflow under the web, while the latter visually displays the final fiber laying effect. The combination of both provides a more comprehensive basis for judging web-forming quality. Through the above-described specific data definitions, precise and multi-dimensional input is provided for subsequent model matching, comparison, and simulation. Alternatively, a laser velocimeter can be used for non-contact measurement of the fiber velocity field.

[0037] Furthermore, the reference process digital twin model is pre-stored in the following manner: Non-Newtonian fluid simulation optimization of the coat hanger type spinning die flow channel was carried out by parametric modeling and genetic algorithm. The coat hanger type flow channel model was established, and the coat hanger type flow channel model defined the mapping relationship between the flow channel structure parameters and the melt flow uniformity. The side-blowing airflow field model is established by simulating the jet inside the side-blowing air box. The side-blowing airflow field model defines the correspondence between the flow guiding structure and the stable flow field pattern. By numerically simulating the interaction between airflow and fiber in the under-net suction system, a coordinated model of under-net suction and wall flow is established. This model defines the control relationship between suction pressure, wall-attached airflow, and fiber web forming uniformity.

[0038] Specifically, the purpose of adopting the above technical solutions is to reveal that the model used as the "judgment standard" is itself a highly specialized and deeply optimized result, rather than a simple data fitting model. For the coat hanger-type flow channel model, its innovation lies in the integration of parametric design, non-Newtonian fluid simulation, and intelligent optimization algorithms, thereby establishing a precise mapping from flow channel geometry to uniform melt distribution at the outlet. For example, during optimization, variables such as manifold inclination angle and slit height may be used, with outlet velocity uniformity as the objective function for optimization. For the fluid model, its core lies in using computational fluid dynamics simulation to pre-calculate a stable flow field without severe pulsations under an ideal guiding structure, and using this as the "gold standard." For the under-net suction and wall flow collaborative model, its innovation lies in studying the complex coupling relationship between multiple airflows, such as suction pressure and wall-attached airflow, and moving fibers through numerical simulation, thereby extracting airflow control rules for achieving stable network formation. The establishment of these models is the foundation for this solution to achieve accurate judgment and optimization, representing a digital accumulation of in-depth process knowledge.

[0039] Furthermore, online simulation and deduction are performed by calling the digital twin model associated with the deviation data, including: When the spinning data deviates, the edge server calls the coat hanger-type flow channel model, and performs rheological calculations based on the real-time melt temperature and melt pressure, and dynamically outputs compensation parameters for the die head temperature zone as the optimization parameters.

[0040] Specifically, the aforementioned technical solution explains how the system uses a model for intelligent diagnosis and parameter compensation when spinning data becomes abnormal. The key lies in "running simplified rheological calculations." This does not involve re-performing a time-consuming full 3D fluid simulation, but rather, based on a pre-stored, validated coat-hanger flow channel model, combined with real-time acquired boundary conditions (melt temperature, pressure), quickly solves a simplified mathematical model. For example, the model might be based on the power-law constitutive equations of non-Newtonian fluids, combined with flow channel geometric parameters, to quickly deduce the equivalent causes of the current uneven pressure distribution, and then calculate the heating power to be compensated or the local resistance coefficient to be adjusted. This optimization parameter is the adjustment amount for the temperature zone setpoint or the fine-tuning mechanism. The threshold here acts as a switch to trigger this simplified simulation process: only when the pressure or temperature deviation exceeds a preset threshold does the system consider model intervention necessary; otherwise, only conventional PID control is performed. This ensures that the system has a low computational load during stable operation, while enabling precise analysis in case of anomalies.

[0041] Furthermore, the online simulation and deduction by invoking the digital twin model associated with the deviation data also includes: When the side-blowing data deviates, the edge server calls the side-blowing flow field model, performs pattern matching between the wind speed distribution and pressure distribution and the pre-stored stable flow field spectrum, and performs rapid simulation of the identified turbulent anomaly region based on the CFD reduced-order model, outputting the adjustment parameters for the fan frequency as the optimization parameters.

[0042] Specifically, the above technical solution aims to automate the diagnosis and adjustment of complex flow field anomalies. Its innovation lies in the combined use of "pattern matching" and "reduced-order computational fluid dynamics (CFD) models." First, the system compares the wind speed / pressure distribution map obtained from the real-time sensor array with the "ideal flow field map" in a pre-existing model (pattern matching) to quickly locate areas with excessive airflow pulsation or uneven distribution. Then, for this abnormal area, a pre-built reduced-order CFD model is invoked. This reduced-order model is a simplified version of the complete CFD model, retaining key features and capable of rapidly simulating the impact of changes in a few control variables, such as guide vane angle and fan speed, on the flow field in that specific area. Through this rapid simulation, the system can evaluate the effectiveness of different adjustment schemes and select the adjustment parameters that best restore the flow field to the "ideal map" state as the optimization command output. This transforms the flow field debugging work, which previously required the intervention of fluid mechanics experts, into an automated data matching and rapid simulation optimization problem.

[0043] Furthermore, the online simulation and deduction by invoking the digital twin model associated with the deviation data also includes: When the negative pressure gradient data or fiber web trajectory image data in the web formation data indicate forming disturbance, the edge server calls the under-web suction wall flow collaborative model, integrates the negative pressure gradient data and fiber trajectory image data to perform gas-solid coupling simulation, and outputs the collaborative adjustment parameters for the main and auxiliary suction air volume and the opening of the attached airflow valve as the optimization parameters.

[0044] Specifically, the aforementioned technical solution addresses the challenge of unstable fiber web formation during high-speed production, with its core innovation lying in data fusion and collaborative control. The system simultaneously monitors the negative pressure gradient (airflow data) and fiber trajectory (visual results). When the image identifies defects such as "web flipping" or "clouding," and the synchronous negative pressure data shows an unreasonable pressure gradient in the corresponding area, model deduction is triggered. The under-web suction-wall flow collaborative model performs a simplified gas-solid coupling simulation, treating the fiber as a group of particles acted upon by multiple airflows (main suction, auxiliary suction, and wall-attached airflow), calculating the fiber's trajectory under the current airflow configuration. Through reverse deduction, the model can determine which airflow parameter caused the observed fiber disturbance. Ultimately, the model outputs not a single adjustment, but a set of collaborative adjustment parameters, such as "increasing the main suction airflow in area A by 5% while reducing the opening of the wall-attached airflow valve in area B by 8%." This multivariate collaborative optimization is difficult to achieve solely through local PID control or manual experience, effectively suppressing crosswind interference and improving web uniformity at high speeds.

[0045] Furthermore, it also includes: The cloud-based remote intelligent control platform receives and aggregates process data, alarm information, and optimization logs from multiple edge servers. The cloud platform maintains an adaptive process formula library based on the received data. The formula library contains a set of preset process parameters for different production scenarios. In response to a scenario selection instruction for a specific production line, the cloud platform retrieves the corresponding set of process parameters from the formula library and sends it to the edge server associated with the target production line to update the corresponding baseline process digital twin model.

[0046] Specifically, the aforementioned technical solution enables centralized management and control of the production network, as well as the accumulation and reuse of knowledge. The cloud platform acts as the "brain," aggregating data and experience from various edge nodes. Its innovation lies in building and maintaining an adaptive process formula library. This library not only statically stores initial process parameter packages for different scenarios but also learns and updates itself based on optimization logs continuously collected from various production lines. For example, a parameter combination successfully optimized for a specific raw material on a production line, after cloud evaluation and verification, can be included in the formula library for other production lines facing similar situations. When a production line needs to switch scenarios, a simple command is issued, and the cloud can remotely deploy the corresponding complete process formula to the line's edge server, enabling it to quickly acquire the production capacity for new products. This significantly reduces the debugging difficulty and knowledge transfer costs of multi-variety, distributed production, realizing the digital assetization of manufacturing experience.

[0047] Furthermore, the cloud-based remote intelligent control platform is also used for: A 3D visualization monitoring interface is provided, which is used to render and display the virtual image of the process status output by the digital twin model of each production line in real time. A remote expert collaboration interface is provided, which allows authorized users to remotely access the augmented reality view of a specific production line and make annotations in the shared view. The annotation information is sent to the corresponding edge server or field terminal as an auxiliary instruction.

[0048] Specifically, the aforementioned technical solutions enhance the intuitiveness of monitoring and the efficiency of remote support. The 3D visualization monitoring interface doesn't simply display data curves; instead, it uses real-time data from each production line to drive their digital twin models, generating a 3D virtual image reflecting the real-time operating status of the equipment, such as a flow cloud map of melt in the die head or a streamline animation of airflow in the bellows. This allows managers to gain a more intuitive and holistic understanding of the production status. The remote expert collaboration interface is an innovative support method. When complex faults occur on-site, personnel can share a first-person view via AR devices. Remote experts, after seeing the synchronized view on the cloud platform, can directly annotate the virtual image, and these annotations are overlaid in real-time on the on-site personnel's AR view. Simultaneously, experts can also use this interface to directly send confirmed debugging commands or temporary model parameter corrections to the edge server. This effectively extends the experts' eyes, brains, and hands remotely to the production site, significantly shortening fault diagnosis and resolution time and providing crucial support for the stable operation of the distributed production network.

[0049] Furthermore, it also includes: The Internet of Things (IoT) continuously collects operational status data of key mechanical components in the production line; Based on the operational status data and historical fault records, the edge server uses a trend analysis algorithm to predict potential component failures and generate predictive maintenance early warning information. The predictive maintenance early warning information is sent to the designated maintenance management terminal.

[0050] Specifically, the aforementioned technical solution expands the control focus from the technological process to the health status of the equipment, shifting from addressing existing problems to preventing future ones. The system continuously collects operational status data characterizing component health, such as the current and vibration spectrum of a screw extruder, the bearing temperature and noise of a blower, and the oil temperature of a gearbox, via the Internet of Things (IoT). The innovation lies in using historical fault records to train or configure trend analysis algorithms and establish component condition degradation models. When real-time data analysis indicates that a certain characteristic parameter exhibits a known pre-fault trend, and its rate of change or absolute value exceeds a safety threshold, the system generates an early warning, such as "The axial vibration amplitude of blower F1 is trending upwards and is expected to exceed the standard within 48 hours; bearing inspection is recommended." This allows maintenance personnel to intervene before a fault occurs, utilizing planned downtime to avoid production losses caused by unplanned downtime and improve overall equipment efficiency.

[0051] Example 2 This invention also discloses a remote intelligent control system for the entire process of nonwoven fabric manufacturing based on the Internet of Things.

[0052] Reference Figure 2 A remote intelligent control system for the entire process of nonwoven fabric manufacturing based on the Internet of Things, including: The equipment layer includes spinning devices, side blowing devices, web forming and suction devices, multiple sensors integrated therein, and controlled actuators; The edge control layer includes an edge server and an IoT gateway deployed locally on the production line. The edge server is used to perform data reception, comparison, simulation and deduction and instruction generation. The IoT gateway is used to perform data communication between the device layer and the edge control layer. The cloud platform layer includes a remote intelligent control platform for cross-production line data aggregation, visualization, formula library management, and remote collaboration support. The edge control layer is connected to the device layer and the cloud platform layer via the Internet of Things, forming a closed loop of data interaction and control.

[0053] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention, they should all fall within the protection scope of the present invention.

[0054] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0055] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A remote intelligent control method for the entire process of nonwoven fabric manufacturing based on the Internet of Things, characterized in that: Includes the following steps: Real-time process data is acquired from multiple sensors deployed on the nonwoven fabric production line; the real-time process data includes spinning data characterizing the melt flow state, side blowing data characterizing the cooling airflow state, and web forming data characterizing the fiber web forming state. The real-time process data is sent to an edge server locally associated with the production line. The edge server matches and compares the received real-time process data with the pre-stored benchmark process digital twin model. The benchmark process digital twin model includes a coat hanger flow channel model based on non-Newtonian fluid simulation, a side-blowing flow field model based on computational fluid dynamics simulation, and a net-mounted suction wall flow collaborative model based on fluid numerical simulation. When the matching comparison result indicates that at least one process data deviates from the preset threshold in the corresponding benchmark process digital twin model, the edge server calls the digital twin model associated with the deviated data to perform online simulation and deduction, so as to determine the process state deviation and obtain the optimization parameters for correcting the deviation. The edge server generates control commands based on the optimized parameters and sends the control commands to the corresponding actuators on the production line via the Internet of Things to adjust the spinning, cooling and web forming processes.

2. The IoT-based remote intelligent control method for the entire process of nonwoven fabric manufacturing, as described in claim 1, is characterized in that... The acquisition of real-time process data from multiple sensors deployed on the nonwoven fabric production line includes: The spinning data is obtained from a sensor embedded in a coat hanger-type spinning die, and the spinning data includes at least melt temperature and melt pressure. The side-blowing data is acquired from a sensor array located on the air outlet surface of the side-blowing box, and the side-blowing data includes at least wind speed distribution and pressure distribution; The web formation data is acquired from a sensor located in the suction duct below the mesh curtain. The web formation data includes at least negative pressure gradient data and fiber web trajectory image data acquired from an image acquisition device facing the web formation area.

3. The method for remote intelligent control of the entire process of nonwoven fabric production based on the Internet of Things as described in claim 1, characterized in that, The baseline process digital twin model is pre-stored in the following manner: Non-Newtonian fluid simulation optimization of the coat hanger type spinning die flow channel was carried out by parametric modeling and genetic algorithm. The coat hanger type flow channel model was established, and the coat hanger type flow channel model defined the mapping relationship between the flow channel structure parameters and the melt flow uniformity. The side-blowing airflow field model is established by simulating the jet inside the side-blowing air box. The side-blowing airflow field model defines the correspondence between the flow guiding structure and the stable flow field pattern. By numerically simulating the interaction between airflow and fiber in the under-net suction system, a coordinated model of under-net suction and wall flow is established. This model defines the control relationship between suction pressure, wall-attached airflow, and fiber web forming uniformity.

4. The IoT-based remote intelligent control method for the entire process of nonwoven fabric manufacturing, as described in claim 3, is characterized in that... Invoking the digital twin model associated with the deviation data for online simulation and deduction includes: When the spinning data deviates, the edge server calls the coat hanger-type flow channel model, and performs rheological calculations based on the real-time melt temperature and melt pressure, and dynamically outputs compensation parameters for the die head temperature zone as the optimization parameters.

5. The IoT-based remote intelligent control method for the entire process of nonwoven fabric manufacturing, as described in claim 3, is characterized in that... The online simulation and deduction process, which involves calling the digital twin model associated with the deviation data, also includes: When the side-blowing data deviates, the edge server calls the side-blowing flow field model, performs pattern matching between the wind speed distribution and pressure distribution and the pre-stored stable flow field spectrum, and performs rapid simulation of the identified turbulent anomaly region based on the CFD reduced-order model, outputting the adjustment parameters for the fan frequency as the optimization parameters.

6. The IoT-based remote intelligent control method for the entire process of nonwoven fabric manufacturing, as described in claim 3, is characterized in that... The online simulation and deduction process, which involves calling the digital twin model associated with the deviation data, also includes: When the negative pressure gradient data or fiber web trajectory image data in the web formation data indicate forming disturbance, the edge server calls the under-web suction wall flow collaborative model, integrates the negative pressure gradient data and fiber trajectory image data to perform gas-solid coupling simulation, and outputs the collaborative adjustment parameters for the main and auxiliary suction air volume and the opening of the attached airflow valve as the optimization parameters.

7. The method for remote intelligent control of the entire process of nonwoven fabric production based on the Internet of Things as described in claim 1, characterized in that, Also includes: The cloud-based remote intelligent control platform receives and aggregates process data, alarm information, and optimization logs from multiple edge servers. The cloud platform maintains an adaptive process formula library based on the received data. The formula library contains a set of preset process parameters for different production scenarios. In response to a scenario selection instruction for a specific production line, the cloud platform retrieves the corresponding set of process parameters from the formula library and sends it to the edge server associated with the target production line to update the corresponding baseline process digital twin model.

8. The method for remote intelligent control of the entire process of nonwoven fabric production based on the Internet of Things as described in claim 1, characterized in that, Also includes: The Internet of Things (IoT) continuously collects operational status data of key mechanical components in the production line; Based on the operational status data and historical fault records, the edge server uses a trend analysis algorithm to predict potential component failures and generate predictive maintenance early warning information. The predictive maintenance early warning information is sent to the designated maintenance management terminal.

9. A remote intelligent control system for the entire process of nonwoven fabric manufacturing based on the Internet of Things (IoT), applied to the remote intelligent control method for the entire process of nonwoven fabric manufacturing based on the IoT described in any one of claims 1-8, characterized in that, include: The equipment layer includes spinning devices, side blowing devices, web forming and suction devices, multiple sensors integrated therein, and controlled actuators; The edge control layer includes an edge server and an IoT gateway deployed locally on the production line. The edge server is used to perform data reception, comparison, simulation and deduction and instruction generation. The IoT gateway is used to perform data communication between the device layer and the edge control layer. The cloud platform layer includes a remote intelligent control platform for cross-production line data aggregation, visualization, formula library management, and remote collaboration support. The edge control layer is connected to the device layer and the cloud platform layer via the Internet of Things, forming a closed loop of data interaction and control.

10. A computer-readable storage medium, characterized in that: The device stores instructions that, when executed on a computer, cause the computer to perform the IoT-based remote intelligent control method for the entire process of nonwoven fabric manufacturing as described in any one of claims 1 to 8.

Images