Die casting machine parameter adaptive optimization method and system based on industrial internet

Abstract

The application discloses a die casting machine parameter self-adaptive optimization method and system based on an industrial internet, relates to the technical field of parameter optimization, and deploys an intelligent gateway and an edge computing unit to collect process parameters, equipment states and quality information of die castings of the die casting machine in real time, drive state synchronization of corresponding digital twins, compare real-time states of the die casting machine with standard states in a process gene library, identify different types of disturbance events, adopt hierarchical collaborative decision-making, generate an optimization strategy sequence of parameters of the die casting machine through knowledge graph reasoning and a federal global optimization model on an industrial internet cloud platform, simulate and verify the optimization strategy by using the digital twins, select an optimal strategy to be converted into equipment control instructions, and issue the equipment control instructions to a target die casting machine for execution. The application solves the problem that existing methods are difficult to adapt to dynamic disturbance events and cannot gather group experience for self-adaptive parameter optimization, and reduces the waste rate caused by fluctuations of the die casting machine.

Description

Technical Field

[0001] This application relates to the field of parameter optimization technology, and in particular to a method and system for adaptive optimization of die-casting machine parameters based on the Industrial Internet. Background Technology

[0002] Die casting, as a high-efficiency and high-precision metal forming process, is widely used in the industrial field. The production quality of die castings fundamentally depends on the precise matching and control of die casting machine parameters in the dynamic process. Finding the optimal combination of process parameters for the die casting machine is the core challenge in the die casting production process.

[0003] Current methods for optimizing die-casting machine parameters often employ empirical trial-and-error, mechanistic models, and traditional machine learning. These methods typically only utilize historical data from a single die-casting machine and cannot leverage collective experience across different equipment or workshops. Consequently, they result in long debugging cycles and significant debugging difficulties when faced with new materials and molds. Existing methods are mostly offline optimization or simple closed-loop control, making it difficult to respond in real time to disturbances such as raw material fluctuations and mold status changes. Furthermore, due to security and privacy considerations, the operating data of die-casting machines in different workshops cannot be aggregated, which restricts optimization capabilities based on big data analysis.

[0004] The Industrial Internet can enable cross-workshop equipment connection of die-casting machines and share operational data. However, the current application of the Industrial Internet focuses on data monitoring and visualization and has not yet penetrated into the real-time adaptive optimization of die-casting machine process parameters. Therefore, there is an urgent need for a die-casting machine parameter adaptive optimization method that can integrate swarm intelligence, virtual environment verification, and real-time response. Summary of the Invention

[0005] To address the technical problems of the prior art, this application provides a method and system for adaptive optimization of die-casting machine parameters based on the Industrial Internet. It constructs a distributed intelligent network based on cloud-edge-device collaboration of the Industrial Internet, and achieves adaptive, traceable, and transferable optimization of die-casting machine parameters through real-time synchronization of digital twins, reasoning of process knowledge graphs, and global modeling of federated learning.

[0006] This application provides an adaptive optimization method for die-casting machine parameters based on the Industrial Internet, including:

[0007] Step S10: Deploy a smart gateway and edge computing unit on each die-casting machine to collect the process parameters and equipment status of the die-casting machine in real time, and communicate with the quality inspection device based on the industrial Internet to obtain the quality information of the die-casting parts, and drive the corresponding unit-level digital twin to achieve status synchronization. In step S20, the digital twin compares the real-time status of each die-casting machine with the standard status in the process gene library, identifies disturbance events related to raw material fluctuations, abnormal mold temperatures, and equipment performance degradation, and assesses the type and severity of the disturbance events. Step S30: Based on the disturbance event evaluation results output by the digital twin, a hierarchical collaborative decision-making approach is adopted. On the industrial internet cloud platform, through knowledge graph reasoning and federated global optimization model, an optimization strategy sequence for die-casting machine parameters is generated and distributed to the workshop edge cloud layer. In step S40, the collaborative optimization agent of the workshop edge cloud layer uses a digital twin to simulate and verify the received optimization strategy in the virtual verification environment, selects the strategy with the best effect and stability, converts it into specific equipment control instructions, and sends them to the target die-casting machine for execution.

[0008] Furthermore, the Industrial Internet consists of the equipment edge layer, the workshop edge cloud layer, and the Industrial Internet cloud platform, which together form a distributed intelligent network with cloud-edge-device collaboration. The equipment edge layer includes intelligent gateways and edge computing units deployed on each die-casting machine. The edge computing units embed a real-time process health assessment module, an edge autonomous control module, and a digital twin drive engine. The workshop edge layer is deployed on the factory's internal servers, including a collaborative optimization agent and a virtual verification environment. It is responsible for managing the collaboration of multiple die-casting machines in the same workshop. By comparing and analyzing the process gene library and the real-time performance of each die-casting machine, it performs safe migration between devices with optimal parameters, synchronously receives optimization strategies from the cloud, and performs high-fidelity simulation verification in the unit-level digital twin to evaluate the effectiveness of the strategies. The industrial internet cloud platform layer is deployed remotely in the cloud and serves as an intelligent hub. It consists of a process knowledge graph module, a federated learning center module, a process gene library management module, and a blockchain evidence storage service module. It enables complex knowledge graph reasoning of the die-casting machine's workflow, updates the federated global optimization model, and achieves model parameter migration across devices and workshops.

[0009] Furthermore, the smart gateway deployed on each die-casting machine collects real-time data at a specific frequency, including die-casting process parameters, die-casting machine status parameters, and quality-related data of the die-cast parts; the edge computing unit receives the real-time data, performs cleaning, alignment, and feature extraction, and drives the corresponding unit-level digital twin to achieve state synchronization; the digital twin includes a physical model of the equipment, a hydraulic system model, and a thermodynamic model, reflecting the virtual state of the physical equipment in real time.

[0010] Furthermore, the edge computing unit, based on the real-time data of the die-casting machine, integrates multiple key performance indicators, analyzes the consistency between the actual injection curve and the set curve, the short-term fluctuation variance of process parameters, calculates the process health index, and triggers an anomaly warning when the index is lower than the dynamic threshold. When the digital twin receives an anomaly warning, it compares the current operating status of the die-casting machine with the benchmark in the process gene library to distinguish whether it is raw material fluctuation, mold status change, equipment performance degradation or environmental interference. Through the built-in parameter causal discovery engine, it analyzes the potential impact of the disturbance on the key quality indicators of the die-casting parts and classifies the disturbance event into local disturbance and global disturbance.

[0011] Furthermore, in response to local disturbances, the autonomous control module embedded in the edge computing unit automatically adjusts the die-casting machine parameters within the production cycle based on a pre-set expert rule base to compensate for fluctuation errors. To address global disturbances, the following collaborative decision-making process is adopted to optimize the adjustment strategy for die-casting machine parameters: The disturbance feature vector provided by the edge node of the die casting machine workshop is obtained, and together with the die casting product model, mold ID and material batch number information, the problem features are generated after desensitization processing, and then packaged into an optimization request and uploaded to the industrial internet cloud platform. The process knowledge graph module in the industrial internet cloud platform receives requests and drives the graph traversal engine to perform similarity matching based on material, mold, and defect entity nodes to find historical successful cases. Through relational path reasoning, it generates a set of preliminary adjustment suggestions for die-casting machine parameters that meet process constraints and retrieves the record with the highest matching degree from the process gene library as a candidate strategy. The Federated Learning Center module takes the problem features and candidate strategies generated from the knowledge graph as input and feeds them into the global federated optimization model. This model has learned the complex die-casting process mapping relationships across factories and materials while protecting the data privacy of all participants. The global federated optimization model evaluates and simulates each candidate strategy, predicts the quality results and production efficiency of the die castings, outputs a sequence of optimized strategies ranked by comprehensive utility, and labels each strategy with confidence level and potential risk points. The ranked optimized strategy sequence and related die casting machine parameter adjustments are then sent to the workshop edge cloud that initiated the request.

[0012] Furthermore, after receiving the strategy sequence, the collaborative optimization agent of the workshop edge cloud performs simulated production in the corresponding unit-level digital twin according to the strategy sequence, verifies the impact of parameter changes on filling, solidification, and stress in the die-casting process, predicts potential defects, selects the optimal strategy that is both effective and stable, and converts the verified strategy into specific equipment control instructions.

[0013] This application also provides an adaptive optimization system for die-casting machine parameters based on the Industrial Internet, including: Real-time sensing and synchronization module: This module is used to deploy intelligent gateways and edge computing units on each die-casting machine to collect process parameters and equipment status of the die-casting machine in real time. It also communicates with quality inspection devices based on the Industrial Internet to obtain quality information of the die-cast parts and drive the corresponding unit-level digital twin to achieve status synchronization. Disturbance identification and assessment module: This module is used by the digital twin to compare the real-time status of each die-casting machine with the standard status in the process gene library, identify disturbance events related to raw material fluctuations, abnormal mold temperatures, and equipment performance degradation, and assess the type and severity of the disturbance events. Layered distributed decision optimization module: Based on the disturbance event evaluation results output by the digital twin, it adopts a layered collaborative decision-making approach, uses knowledge graph reasoning and federated global optimization model on the industrial internet cloud platform to generate an optimization strategy sequence for die-casting machine parameters, and distributes it to the edge cloud layer of the workshop; Virtual Verification and Execution Module: The collaborative optimization agent for the workshop edge cloud layer uses a digital twin to simulate and verify the received optimization strategies in the virtual verification environment, selects the most effective and stable strategy, converts it into specific equipment control commands, and sends them to the target die-casting machine for execution.

[0014] This application discloses the following technical effects: This application provides a method and system for adaptive optimization of die-casting machine parameters based on the Industrial Internet. Relying on the Industrial Internet platform, it automatically responds to uncertainties and disturbances during the operation of the die-casting machine through a closed loop of real-time perception, digital twin analysis, cloud-edge collaborative decision-making, and virtual verification. It identifies and compensates for process drift before or in its early stages, significantly reducing the scrap rate caused by fluctuations in the die-casting machine equipment. This method uses knowledge graphs and federated learning models to recommend die-casting machine parameters, combined with virtual trial and error through digital twins, to shorten the process debugging time for new molds and new products. Through adaptive optimization, it keeps the die-casting machine running continuously under optimal parameters, improving equipment utilization and capacity. In addition, the federated learning mechanism in this method aggregates die-casting machine operating status data and process parameters from different workshops, factories, and even enterprises to jointly train a more powerful and robust global optimization model, realizing the coexistence of group experience and breaking the constraints of data silos on the performance improvement of intelligent models. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0016] Figure 1 This is a flowchart illustrating an adaptive optimization method for die-casting machine parameters based on the Industrial Internet, provided in an embodiment of this application.

[0017] Figure 2 This is a schematic diagram of a die-casting machine parameter adaptive optimization system based on the Industrial Internet, provided in an embodiment of this application. Detailed Implementation

[0018] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0019] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] In the following description, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those explicitly listed, but may include other steps or modules not explicitly listed or inherent to such processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0021] Example 1: This application provides an adaptive optimization method for die-casting machine parameters based on the Industrial Internet. The Industrial Internet consists of an equipment edge layer, a workshop edge cloud layer, and an Industrial Internet cloud platform, which together form a distributed intelligent network with cloud-edge-device collaboration. The device edge layer includes smart gateways and edge computing units deployed on each die-casting machine, wherein the edge computing units embed the following modules: Real-time process health assessment module: Calculates the process health index of the die-casting cycle based on real-time data; Edge Autonomous Control Module: Built-in local rule base, performs millisecond-level parameter fine-tuning for predictable disturbances; Digital Twin Drive Engine: Drives and maintains a unit-level digital twin that is synchronized in real time with the physical die-casting machine; the twin contains a parameter causality discovery submodule for dynamically analyzing the causal relationship between parameters and quality indicators; The workshop edge layer is deployed on the factory's internal servers, including a collaborative optimization agent and a virtual verification environment. It is responsible for managing the collaboration of multiple die-casting machines in the same workshop. By comparing and analyzing the process gene library and the real-time performance of each die-casting machine, it performs safe migration between devices with optimal parameters, synchronously receives optimization strategies from the cloud, and performs high-fidelity simulation verification in the unit-level digital twin to evaluate the effectiveness of the strategies. The industrial internet cloud platform layer is deployed remotely in the cloud and serves as an intelligent hub, consisting of the following modules: Process knowledge graph module: Stores and manages a network of knowledge consisting of a material library, a mold library, a historical process library, and a defect pattern library, supporting semantic-based queries and reasoning of complex die-casting processes; Federated Learning Center Module: Includes a global optimization model and a federated scheduler. The federated scheduler coordinates the edge nodes of each workshop, enabling them to train local models using local data and only upload encrypted model gradients or parameter updates. The global optimization model is then securely aggregated and updated at the center, and the enhanced model is then distributed to each node. Process gene library management module: Standardizes and encapsulates validated optimization parameter sets, generates process gene packages with operating condition tags, and distributes them across workshops and factories within the authorized scope; Blockchain Evidence Preservation Service Module: Generates tamper-proof evidence records for each key parameter adjustment, optimization decision trigger, and corresponding quality result, forming trusted evidence preservation.

[0022] like Figure 1 As shown, the adaptive optimization method for die-casting machine parameters based on the Industrial Internet includes: Step S10: Deploy a smart gateway and edge computing unit on each die-casting machine to collect the process parameters and equipment status of the die-casting machine in real time, and communicate with the quality inspection device based on the industrial Internet to obtain the quality information of the die-cast parts, and drive the corresponding unit-level digital twin to achieve status synchronization.

[0023] In this embodiment, the intelligent gateway collects the process parameters and equipment status of the die-casting machine in real time, including at least: Process parameters: displacement and velocity curve of injection punch, injection pressure curve, boosting pressure and pressure build-up time, barrel temperature, temperature at various points of the mold, cooling water flow rate and temperature, and vacuum degree; Status parameters: hydraulic system oil temperature, valve response signals, and equipment operating cycle time; Meanwhile, the edge computing unit communicates with the server of the die-casting quality inspection device (industrial vision inspection system, in-mold pressure sensor array, and X-ray online inspection unit) via OPC UA, MQTT protocol, or dedicated API to acquire and correlate quality data. The detailed steps are as follows: When the die-casting machine completes a cycle and the die-cast part is sent to the inspection station, the inspection device reads the unique identifier (RFID or QR code, which has been associated with process data during the die-casting process) on the casting. The detection device correlates the detection results—namely, the flash, scratches, cold shut defects and their locations derived from image analysis, the filling integrity derived from in-mold pressure curve analysis, and the porosity level determined by X-ray—with the die casting to generate a quality data package. This quality data package is then pushed and acquired in real time through the industrial internet network, ensuring that the quality results of the die casting are accurately correlated with the process data of the specific die casting cycle that produced it. Secondly, all collected data are preprocessed to remove abnormal jump points, align data from different subsystems based on a unified time scale, and calculate higher-order features, including at least peak injection velocity, average velocity, pressure rise gradient, mold temperature difference, and cycle time stability, to reduce the amount of data. Next, the process parameters, equipment status, and associated quality data packets generated for each die-casting cycle are used to drive the unit-level digital twin corresponding to the physical die-casting machine. This digital twin includes a physical model of the equipment, a hydraulic system model, and a thermodynamic model, reflecting the virtual state of the physical equipment in real time. Real-time data is injected into the digital twin model as boundary conditions and inputs, including at least: The actual injection speed curve is used as input to drive the movement of the punch in the hydraulic system model; the actual mold temperature measurement data is used as the reverse input to the thermodynamic model to calibrate and update the temperature field of the entire virtual mold. After receiving real-time data, the digital twin mirrors the current state of the physical die-casting machine in virtual space; operators can see the virtual die-casting machine running at a speed synchronized with the physical world through the monitoring interface, including valve core movement, injection process animation, and temperature field cloud map changes.

[0024] In step S20, the digital twin compares the real-time status of each die-casting machine with the standard status in the process gene library, identifies disturbance events related to raw material fluctuations, abnormal mold temperatures, and equipment performance degradation, and assesses the type and severity of the disturbance events.

[0025] In this embodiment, the edge computing unit, based on the real-time data of the die-casting machine, integrates multiple key performance indicators, analyzes the consistency between the actual injection curve and the set curve, the short-term fluctuation variance of the process parameters, calculates the process health index, and triggers an abnormal warning when the index is lower than the dynamic threshold. When the digital twin receives an anomaly warning, it compares the current operating status of the die-casting machine with the benchmark in the process gene library to distinguish whether it is a fluctuation in raw materials, a change in mold status, or a decline in equipment performance. The process gene library stores standard production states corresponding to different die-casting products, molds, and materials, providing multi-dimensional dynamic reference benchmarks, including at least: Gold die casting process curve: ideal standard injection speed and pressure curve; Standard thermal equilibrium field: Standard temperature values ​​and allowable fluctuation ranges at key points of the mold; Quality-related thresholds: target values ​​and acceptable ranges for key process characteristics of die-cast parts; Equipment condition baseline: normal hydraulic system response and cycle time; The digital twin performs multi-level, multi-index comparison and calculation between the real-time status of each die-casting machine and the standard status in the process gene library, including at least the following detailed steps: Process consistency comparison: Calculate the cosine similarity between the real-time injection speed and pressure curve of the die-casting machine and the gold die-casting process curve. When the similarity curve shows shape distortion or peak lag, it is determined that there is a problem of die-casting machine equipment attenuation or hydraulic system abnormality. Thermal equilibrium state comparison: The real-time acquired multi-point temperature sequence of the die-casting machine mold is compared with the standard thermal equilibrium length to analyze the distribution pattern of the entire temperature field. It is then matched with a predefined mold temperature anomaly pattern library to determine whether the following anomaly patterns exist: Localized overheating mode: The temperature at a certain point is consistently higher than the standard and the gradient is abnormal, indicating that the mold cooling water channel is blocked; Overall drift mode: The temperature at all measuring points rises or falls in the same direction, which may be caused by changes in ambient temperature or total cooling water temperature. Gradient reversal mode: The temperature gradient is opposite to the design direction, indicating a fundamental change in the internal heat conduction state of the mold, such as coating peeling or cracking. Quality-related parameter deviation analysis: Based on real-time process data, the predicted porosity and shrinkage risk index of the die casting in the current cycle are calculated through the built-in filling simulation prediction. The predicted virtual quality indicators are compared with the qualified standards defined in the gene library. If the predicted risk is higher than the threshold, it is marked as a raw material fluctuation event.

[0026] The built-in parameter causal discovery engine analyzes the potential impact of disturbances on key quality indicators of die-cast parts, classifying disturbance events into local and global disturbances. Local disturbances: Disturbances caused by a single factor, with a small range of impact and rapid changes. Such disturbances can usually be quickly compensated for by preset rules. Global disturbances: disturbances caused by fundamental changes, with a wide impact range, or involving multi-parameter coupling; for example, changes in the performance of new batches of raw materials, long-term thermal equilibrium drift caused by mold wear, and slow system response caused by long-term deterioration of hydraulic oil; such disturbances require systematic optimization based on cloud-edge collaborative decision-making.

[0027] Step S30: Based on the disturbance event evaluation results output by the digital twin, a hierarchical collaborative decision-making approach is adopted. On the industrial internet cloud platform, through knowledge graph reasoning and federated global optimization model, an optimization strategy sequence for die-casting machine parameters is generated and distributed to the edge cloud layer of the workshop.

[0028] In this embodiment, for local disturbances, the autonomous control module embedded in the edge computing unit automatically adjusts the die-casting machine parameters within the production cycle according to the preset expert rule base to compensate for fluctuation errors. To address global disturbances, the following collaborative decision-making process is adopted to optimize the adjustment strategy for die-casting machine parameters: The disturbance feature vector provided by the edge node of the die casting machine workshop is obtained, and together with the die casting product model, mold ID and material batch number information, the problem features are generated after desensitization processing, and then packaged into an optimization request and uploaded to the industrial internet cloud platform. The process knowledge graph module in the industrial internet cloud platform receives requests and drives the graph traversal engine to perform similarity matching based on material, mold, and defect entity nodes to find historical successful cases. Through relational path reasoning, it generates a set of preliminary adjustment suggestions for die-casting machine parameters that meet process constraints and retrieves the record with the highest matching degree from the process gene library as a candidate strategy. The Federated Learning Center module takes the problem features and candidate strategies generated from the knowledge graph as input and feeds them into the global federated optimization model. This model has learned the complex die-casting process mapping relationships across factories and materials while protecting the data privacy of all participants. The global federated optimization model evaluates and simulates each candidate strategy, predicts the quality results and production efficiency of the die castings, outputs a sequence of optimized strategies ranked by comprehensive utility, and labels each strategy with confidence level and potential risk points. The ranked optimized strategy sequence and related die casting machine parameter adjustments are then sent to the workshop edge cloud that initiated the request.

[0029] In step S40, the collaborative optimization agent of the workshop edge cloud layer uses a digital twin to simulate and verify the received optimization strategy in the virtual verification environment, selects the strategy with the best effect and stability, converts it into specific equipment control instructions, and sends them to the target die-casting machine for execution.

[0030] In this embodiment, after receiving the strategy sequence, the collaborative optimization agent of the workshop edge cloud layer simulates production in the corresponding unit-level digital twin according to the strategy sequence, verifies the impact of parameter changes on filling, solidification and stress in the die casting process, predicts potential defects, selects the optimal strategy that is both effective and stable, and converts the verified strategy into specific equipment control instructions. The actual data and quality results of the die-casting machine after the strategy is executed are fed back to the edge nodes, and successful optimization cases are added to the process gene library. Their metadata, decision logic, and result hashes are synchronized to the cloud platform. Specific knowledge about die-casting machine parameter optimization is stored in the process knowledge graph to enhance reasoning ability; process data of die-casting parameter optimization is used for the next round of model training in federated learning to improve global intelligence; key logs are stored in the blockchain after hashing to complete trusted evidence storage.

[0031] Example 2: The adaptive optimization system for die-casting machine parameters based on the Industrial Internet provided in this embodiment of the invention can execute the adaptive optimization method for die-casting machine parameters based on the Industrial Internet provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method, such as... Figure 2 As shown, it includes the following modules: Real-time sensing and synchronization module: This module is used to deploy intelligent gateways and edge computing units on each die-casting machine to collect process parameters and equipment status of the die-casting machine in real time. It also communicates with quality inspection devices based on the Industrial Internet to obtain quality information of the die-cast parts and drive the corresponding unit-level digital twin to achieve status synchronization. Disturbance identification and assessment module: This module is used by the digital twin to compare the real-time status of each die-casting machine with the standard status in the process gene library, identify disturbance events related to raw material fluctuations, abnormal mold temperatures, and equipment performance degradation, and assess the type and severity of the disturbance events. Layered distributed decision optimization module: Based on the disturbance event evaluation results output by the digital twin, it adopts a layered collaborative decision-making approach, uses knowledge graph reasoning and federated global optimization model on the industrial internet cloud platform to generate an optimization strategy sequence for die-casting machine parameters, and distributes it to the edge cloud layer of the workshop; Virtual Verification and Execution Module: The collaborative optimization agent for the workshop edge cloud layer uses a digital twin to simulate and verify the received optimization strategies in the virtual verification environment, selects the most effective and stable strategy, converts it into specific equipment control commands, and sends them to the target die-casting machine for execution.

[0032] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0033] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. An adaptive optimization method for die-casting machine parameters based on the Industrial Internet, characterized in that, The method includes: Step S10: Deploy a smart gateway and edge computing unit on each die-casting machine to collect the process parameters and equipment status of the die-casting machine in real time, and communicate with the quality inspection device based on the industrial Internet to obtain the quality information of the die-casting parts, and drive the corresponding unit-level digital twin to achieve status synchronization. In step S20, the digital twin compares the real-time status of each die-casting machine with the standard status in the process gene library, identifies disturbance events related to raw material fluctuations, abnormal mold temperatures, and equipment performance degradation, and assesses the type and severity of the disturbance events. Step S30: Based on the disturbance event evaluation results output by the digital twin, a hierarchical collaborative decision-making approach is adopted. On the industrial internet cloud platform, through knowledge graph reasoning and federated global optimization model, an optimization strategy sequence for die-casting machine parameters is generated and distributed to the workshop edge cloud layer. In step S40, the collaborative optimization agent of the workshop edge cloud layer uses a digital twin to simulate and verify the received optimization strategy in the virtual verification environment, selects the strategy with the best effect and stability, converts it into specific equipment control instructions, and sends them to the target die-casting machine for execution.

2. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 1, characterized in that, The industrial internet consists of a device edge layer, a workshop edge cloud layer, and an industrial internet cloud platform, which together form a distributed intelligent network with cloud-edge-device collaboration. The equipment edge layer includes intelligent gateways and edge computing units deployed on each die-casting machine. The edge computing unit embeds a real-time process health assessment module, an edge autonomous control module, and a digital twin drive engine. The workshop edge layer is deployed on the factory's internal servers, including a collaborative optimization agent and a virtual verification environment, and is responsible for managing the collaboration of multiple die-casting machines in the same workshop; By comparing and analyzing the process gene library and the real-time performance of each die-casting machine, safe migration between devices with optimal parameters is performed, and the effectiveness of the strategy is evaluated simultaneously. The industrial internet cloud platform layer is deployed remotely in the cloud and serves as an intelligent hub. It consists of a process knowledge graph module, a federated learning center module, a process gene library management module, and a blockchain evidence storage service module. It enables complex knowledge graph reasoning of the die-casting machine's workflow, updates the federated global optimization model, and completes the migration of model parameters across devices and workshops.

3. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 1, characterized in that, In step S10, the smart gateway collects real-time data of the die-casting machine at a specific frequency, including die-casting process parameters, die-casting machine status parameters, and quality-related data of the die-cast parts; The edge computing unit receives real-time data, performs cleaning, alignment, and feature extraction, and drives the corresponding unit-level digital twin to achieve state synchronization. The digital twin includes a physical model of the equipment, a hydraulic system model, and a thermodynamic model, reflecting the virtual state of the physical die-casting machine in real time.

4. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 3, characterized in that, The acquisition of the quality correlation data includes the following detailed steps: When the die-casting machine completes a cycle and the die-cast part is sent to the inspection station, the inspection device reads the unique identifier on the die-cast part and associates the inspection results—namely, the flash, scratches, cold shut defects and their locations obtained from image analysis, the filling integrity obtained from in-mold pressure curve analysis, and the porosity level determined by X-ray—with the corresponding die-cast part to generate a quality data package. This quality data package is obtained in real time through the industrial internet network, ensuring that the quality results of the die-cast part are accurately associated with the process data of the specific die-casting cycle that produced it.

5. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 1, characterized in that, In step S20, the process gene library stores standard production states corresponding to different die-casting products, molds, and materials, providing multi-dimensional dynamic reference benchmarks, including: Gold die casting process curve: ideal standard injection speed and pressure curve; Standard thermal equilibrium field: Standard temperature values ​​and allowable fluctuation ranges at key points of the mold; Quality-related thresholds: target values ​​and acceptable ranges for key process characteristics of die-cast parts; Equipment status baseline: normal hydraulic system response and cycle time.

6. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 1, characterized in that, In step S20, the digital twin performs multi-level, multi-index comparison calculations between the real-time status of each die-casting machine and the standard status in the process gene library, including the following detailed steps: Process consistency comparison: Calculate the cosine similarity between the real-time injection speed and pressure curve of the die-casting machine and the gold die-casting process curve. When the similarity curve shows shape distortion or peak lag, it is determined that there is a problem of die-casting machine equipment attenuation or hydraulic system abnormality. Thermal equilibrium state comparison: The real-time collected multi-point temperature sequence of the die-casting machine mold is compared with the standard thermal equilibrium field to analyze the distribution pattern of the entire temperature field and match it with the predefined mold temperature anomaly pattern library to determine whether there are local overheating patterns, overall drift patterns and gradient reversal patterns. Quality-related parameter deviation analysis: Based on real-time process data, the predicted porosity and shrinkage risk index of the die casting in the current cycle are calculated through the built-in filling simulation prediction. The predicted virtual quality indicators are compared with the qualified standards defined in the gene library. If the predicted risk is higher than the threshold, it is marked as a raw material fluctuation event.

7. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 1, characterized in that, In step S30, for local disturbance events in the digital twin evaluation, the autonomous control module embedded in the edge computing unit automatically adjusts the die-casting machine parameters within the production cycle according to the preset expert rule base to compensate for fluctuation errors; for global disturbance events in the digital twin evaluation, a cloud-edge collaborative decision-making process is adopted to optimize the adjustment strategy of the die-casting machine parameters.

8. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 7, characterized in that, The cloud-edge collaborative decision-making process includes: Obtain the disturbance feature vector provided by the edge node of the die casting machine workshop, and together with the die casting product model, mold ID and material batch number information, generate problem features and optimization requests, and upload them to the industrial internet cloud platform; The process knowledge graph module in the industrial internet cloud platform receives requests and drives the graph traversal engine to perform similarity matching based on material, mold, and defect entity nodes to find historical successful cases; through relational path reasoning, it generates preliminary adjustment suggestions for die-casting machine parameters and retrieves the record with the highest matching degree from the process gene library as a candidate strategy; The Federated Learning Center module takes problem characteristics and candidate strategies as input and feeds them into the global federated optimization model, which has learned complex die-casting process mapping relationships across plants and materials. The global federated optimization model evaluates and simulates each candidate strategy, predicts the quality results and production efficiency of the die castings, and outputs a sequence of optimized strategies ranked by comprehensive utility. The ranked optimized strategy sequence and related die casting machine parameter adjustments are then sent to the workshop edge cloud that initiated the request.

9. The adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in claim 1, characterized in that, In step S40, after receiving the strategy sequence, the collaborative optimization agent of the workshop edge cloud layer simulates production in the corresponding unit-level digital twin according to the strategy sequence, verifies the impact of parameter changes on filling, solidification and stress in the die casting process, predicts potential defects, selects the optimal strategy that is both effective and stable, and converts the verified strategy into specific equipment control instructions. The actual data and quality results of the die-casting machine after the strategy is executed are fed back to the edge node. Successful optimization cases are added to the process gene library. Specific knowledge about die-casting machine parameter optimization is stored in the process knowledge graph. The process data of die-casting parameter optimization is used for the next round of model training in federated learning. Key log records are stored in the blockchain after hashing to complete trusted evidence storage.

10. An adaptive optimization system for die-casting machine parameters based on the Industrial Internet, characterized in that, The system is used to implement the adaptive optimization method for die-casting machine parameters based on the Industrial Internet as described in any one of claims 1-9, and the system includes: Real-time sensing and synchronization module: This module is used to deploy intelligent gateways and edge computing units on each die-casting machine to collect process parameters and equipment status of the die-casting machine in real time. It also communicates with quality inspection devices based on the Industrial Internet to obtain quality information of the die-cast parts and drive the corresponding unit-level digital twin to achieve status synchronization. Disturbance identification and assessment module: This module is used by the digital twin to compare the real-time status of each die-casting machine with the standard status in the process gene library, identify disturbance events related to raw material fluctuations, abnormal mold temperatures, and equipment performance degradation, and assess the type and severity of the disturbance events. Layered distributed decision optimization module: Based on the disturbance event evaluation results output by the digital twin, it adopts a layered collaborative decision-making approach, uses knowledge graph reasoning and federated global optimization model on the industrial internet cloud platform to generate an optimization strategy sequence for die-casting machine parameters, and distributes it to the edge cloud layer of the workshop; Virtual Verification and Execution Module: The collaborative optimization agent for the workshop edge cloud layer uses a digital twin to simulate and verify the received optimization strategies in the virtual verification environment, selects the most effective and stable strategy, converts it into specific equipment control commands, and sends them to the target die-casting machine for execution.

Images