Extruder screw rotating speed intelligent matching and energy consumption compensation method based on deep learning

Abstract

The invention discloses an extruder screw rotating speed intelligent matching and energy consumption compensation method based on deep learning, and relates to the technical field of monitoring analysis, and the method comprises the steps: collecting operation state parameters and material characteristic parameters corresponding to an extruder, constructing a material-equipment coupling data set, carrying out screw equivalent torque analysis based on the material-equipment coupling data set, and carrying out screw equivalent torque analysis based on the material-equipment coupling data set. Generating a screw load characteristic spectrum; temperature-pressure coupling field evolution simulation is carried out according to the screw load characteristic spectrum, melt phase change process parameters are obtained, and then an intelligent rotating speed control curve is generated; establishing an energy consumption compensation coefficient matrix, and performing phase difference compensation mapping on the energy consumption compensation coefficient matrix according to the electric parameter characteristics of the extruder driving motor winding to generate an energy consumption compensation instruction set; and constructing an energy consumption dynamic balance model, carrying out online optimization on the energy consumption dynamic balance model, generating a screw rotation speed-energy consumption cooperative controller, and deploying the screw rotation speed-energy consumption cooperative controller to an edge computing terminal.

Description

Technical Field

[0001] This application relates to the field of monitoring and analysis technology, and in particular to a method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning. Background Technology

[0002] An extruder, also known as an extruder, is a type of wire and cable equipment mainly used for extruding the core wires of various wires and cables, such as power cords, power cables, coaxial cables, communication cables, transmission cables, BV-class civilian wires, electronic wires, computer wires, building wires, data communication cables, radio frequency wires, HDMI data cables, various high-temperature silicone cables, Teflon cables, coaxial cables, network cables, optical fibers, and drop optical cables.

[0003] In related technologies, traditional screw speed regulation mainly relies on operators' experience to set fixed parameters. As a result, when material properties or environmental conditions change, the fixed speed mode leads to aggravated melt pressure pulsation, causing a sharp increase in energy consumption and material degradation. Moreover, current energy compensation mostly adopts a feedback voltage regulation strategy, without establishing a real-time correlation model between the drive motor and screw load fluctuations. This results in significant lag in energy compensation under transient operating conditions, thereby reducing energy compensation efficiency, which needs to be improved. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides a deep learning-based method for intelligent matching of extruder screw speed and energy consumption compensation.

[0005] In the first aspect, this application provides a deep learning-based method for intelligent matching of extruder screw speed and energy consumption compensation, including the following steps:

[0006] Step S1: Collect the operating status parameters and material characteristic parameters corresponding to the extruder, construct a material-equipment coupled dataset, perform screw equivalent torque analysis based on the material-equipment coupled dataset, and generate a screw load characteristic map;

[0007] Step S2: Perform temperature-pressure coupled field evolution simulation based on screw load characteristic spectrum to obtain melt phase transformation process parameters, perform screw speed matching degree iterative calculation on melt phase transformation process parameters, and generate intelligent speed control curve;

[0008] Step S3: Perform dynamic compensation analysis on energy consumption fluctuations of the intelligent speed control curve, establish an energy consumption compensation coefficient matrix, perform phase difference compensation mapping on the energy consumption compensation coefficient matrix based on the electrical parameter characteristics of the extruder drive motor winding, and generate an energy consumption compensation instruction set.

[0009] Step S4: Construct an energy consumption dynamic balance model based on the intelligent speed control curve and energy consumption compensation instruction set, optimize the energy consumption dynamic balance model online, generate a screw speed-energy consumption co-controller, and deploy it to the edge computing terminal.

[0010] Preferably, step S1 includes the following steps:

[0011] The axial distributed pressure value corresponding to the extruder screw is collected by a pressure sensor array, and the melt index and thermal conductivity parameters of the material are obtained by an infrared spectroscopy sensor. Then, a material-equipment coupled dataset is constructed based on the axial distributed pressure value, the melt index and thermal conductivity parameters of the material.

[0012] The material-equipment coupled dataset is denoised, and a denoised feature vector is generated based on the denoising result. An equivalent mechanical model is constructed, and the denoised feature vector is input into the equivalent mechanical model for circumferential shear stress analysis. Load features are extracted from the circumferential shear stress analysis results to generate a screw load feature map.

[0013] Preferably, step S2 includes the following steps:

[0014] The dynamic load characteristic spectrum of the screw is divided into multiple functional segments, including melt conveying functional segment, compression functional segment and metering functional segment, and the temperature field gradient and pressure field gradient correlation matrix corresponding to each functional segment is established.

[0015] Dynamic evolution simulation is performed on the correlation matrix of temperature field gradient and pressure field gradient corresponding to each functional segment, and the melt phase change process parameters corresponding to each functional segment are obtained based on the results of the dynamic evolution simulation. The melt phase change process parameters include melt phase change rate and viscosity distribution parameters.

[0016] A state-space equation is constructed, with the melt phase transformation rate and viscosity distribution parameters as input constraints. The optimal speed set is then solved by interpolating the optimal speed set to generate an intelligent speed control curve.

[0017] Preferably, step S3 includes the following steps:

[0018] The electrical parameter characteristics corresponding to the operation of the extruder drive motor winding are collected. The electrical parameter characteristics include three-phase current ripple characteristics and harmonic distortion. The current-speed correlation feature surface is constructed by combining the intelligent speed control curve.

[0019] The current-speed correlation feature surface is subjected to fluctuation decomposition, and then the load mutation data and transient energy overshoot region are identified based on the results of the fluctuation decomposition. The load mutation data includes the load mutation point and the load mutation time.

[0020] An energy consumption compensation coefficient matrix is ​​established. Based on the energy consumption compensation coefficient matrix, a zero-phase difference filter bank is configured for the transient energy overshoot region. The corresponding compensation coefficient combination of the filter bank is optimized to generate an energy consumption compensation instruction set.

[0021] Preferably, after generating the energy consumption compensation instruction set, it also includes:

[0022] A compensation priority evaluation model is established based on load mutation data, and then the energy consumption compensation instruction set is dynamically prioritized based on the compensation priority evaluation model.

[0023] The edge computing terminal monitors and provides feedback on the compensation effect of the energy consumption compensation instruction set in real time, and then optimizes the energy consumption compensation instruction based on the monitoring and feedback results.

[0024] Preferably, step S4 includes the following steps:

[0025] Construct a target optimization function, input the intelligent speed control curve and energy consumption compensation instruction set as constraints into the target optimization function, perform balance training on the target optimization function, and generate a dynamic energy consumption balance model.

[0026] A virtual extruder simulation environment is established through a digital twin platform. The energy consumption dynamic balance model is then subjected to transfer reinforcement learning using a set of historical process parameters. The energy consumption dynamic balance model after transfer reinforcement learning is encapsulated into a screw speed-energy consumption co-controller, and the screw speed-energy consumption co-controller is deployed to an edge computing terminal.

[0027] Preferably, after deploying the screw speed-energy consumption co-controller to the edge computing terminal, it further includes:

[0028] Within a preset period, the screw speed-energy consumption co-controller is remotely monitored, and the corresponding optimization data of the screw speed-energy consumption co-controller is collected in real time. Based on the optimization data, the parameters of the screw speed-energy consumption co-controller are tuned.

[0029] Secondly, this application provides a deep learning-based intelligent matching and energy consumption compensation system for extruder screw speed, including:

[0030] The data acquisition module is used to collect the operating status parameters and material characteristic parameters of the extruder, construct a material-equipment coupled dataset, perform screw equivalent torque analysis based on the material-equipment coupled dataset, and generate a screw load characteristic map.

[0031] The analysis module is used to simulate the evolution of the temperature-pressure coupled field based on the screw load characteristic spectrum, obtain the melt phase change process parameters, perform iterative calculation of the screw speed matching degree on the melt phase change process parameters, and generate intelligent speed control curves.

[0032] The instruction generation module is used to perform dynamic compensation analysis of energy consumption fluctuations on the intelligent speed control curve, establish an energy consumption compensation coefficient matrix, perform phase difference compensation mapping on the energy consumption compensation coefficient matrix based on the electrical parameter characteristics of the extruder drive motor winding, and generate an energy consumption compensation instruction set.

[0033] The optimization module is used to build an energy consumption dynamic balance model based on the intelligent speed control curve and energy consumption compensation instruction set, optimize the energy consumption dynamic balance model online, generate a screw speed-energy consumption co-controller, and deploy it to the edge computing terminal.

[0034] Thirdly, this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the deep learning-based intelligent matching and energy consumption compensation method for extruder screw speed described in any of the above-mentioned embodiments.

[0035] In summary, this application includes the following beneficial technical effects:

[0036] This application provides a deep learning-based intelligent screw speed matching and energy consumption compensation method for extruders. By analyzing the operating state parameters and material characteristic parameters of the extruder, a screw load characteristic map is generated. Based on this map, a temperature-pressure coupled field evolution simulation is performed to obtain melt phase change process parameters. The screw speed matching degree is iteratively calculated on these parameters to generate an intelligent speed control curve. This reduces the likelihood of screw speed adjustment relying primarily on operator experience and fixed parameters, which can lead to increased melt pressure pulsation, resulting in a sharp increase in energy consumption and material degradation. An energy consumption compensation coefficient matrix is ​​established, and phase difference compensation mapping is performed on the energy consumption compensation coefficient matrix according to the electrical parameter characteristics of the extruder drive motor winding to generate an energy consumption compensation instruction set. An energy consumption dynamic balance model is constructed based on the intelligent speed control curve and the energy consumption compensation instruction set. The energy consumption dynamic balance model is optimized online to generate a screw speed-energy consumption co-controller and deploy it to the edge computing terminal. This effectively establishes a real-time correlation model between the drive motor and the screw load fluctuation, thereby effectively reducing the occurrence of significant energy consumption compensation lag under transient operating conditions and thus effectively improving energy consumption compensation efficiency. Attached Figure Description

[0037] 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.

[0038] Figure 1This is a flowchart of a method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning, according to an embodiment of this application.

[0039] Figure 2 This is a schematic diagram of a deep learning-based intelligent matching and energy consumption compensation system for extruder screw speed in an embodiment of this application. Detailed Implementation

[0040] The following is in conjunction with the appendix Figure 1-2 This application will be described in further detail.

[0041] Example 1

[0042] This application discloses a method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning.

[0043] Reference Figure 1 A deep learning-based intelligent matching and energy consumption compensation method for extruder screw speed includes the following steps:

[0044] Step S1: Collect the operating status parameters and material characteristic parameters corresponding to the extruder, construct a material-equipment coupled dataset, perform screw equivalent torque analysis based on the material-equipment coupled dataset, and generate a screw load characteristic map;

[0045] Step S2: Perform temperature-pressure coupled field evolution simulation based on screw load characteristic spectrum to obtain melt phase transformation process parameters, perform screw speed matching degree iterative calculation on melt phase transformation process parameters, and generate intelligent speed control curve;

[0046] Step S3: Perform dynamic compensation analysis on energy consumption fluctuations of the intelligent speed control curve, establish an energy consumption compensation coefficient matrix, perform phase difference compensation mapping on the energy consumption compensation coefficient matrix based on the electrical parameter characteristics of the extruder drive motor winding, and generate an energy consumption compensation instruction set.

[0047] Step S4: Construct an energy consumption dynamic balance model based on the intelligent speed control curve and energy consumption compensation instruction set, optimize the energy consumption dynamic balance model online, generate a screw speed-energy consumption co-controller, and deploy it to the edge computing terminal.

[0048] It should be noted that step S1 includes the following steps:

[0049] The axial distributed pressure value corresponding to the extruder screw is collected by a pressure sensor array, and the melt index and thermal conductivity parameters of the material are obtained by an infrared spectroscopy sensor. Then, a material-equipment coupled dataset is constructed based on the axial distributed pressure value, the melt index and thermal conductivity parameters of the material.

[0050] The material-equipment coupled dataset is denoised, and a denoised feature vector is generated based on the denoising result. An equivalent mechanical model is constructed, and the denoised feature vector is input into the equivalent mechanical model for circumferential shear stress analysis. Load features are extracted from the circumferential shear stress analysis results to generate a screw load feature map.

[0051] In this embodiment, pressure sensor arrays are arranged in different axial sections of the extruder screw barrel to collect axial pressure values ​​from the feed section to the die head in real time. Simultaneously, the molten material is detected online using an infrared spectroscopy sensor. The melt index of the material is calculated using the characteristic peak intensity, and the thermal conductivity parameter is inverted based on the molecular vibrational energy level transition law. The melt index and thermal conductivity parameter are matched by timestamp to construct a material-equipment coupled dataset. Then, wavelet threshold denoising is used to process the material-equipment coupled dataset to generate a denoised feature vector containing temperature, pressure, and material parameters. An equivalent mechanical model is constructed based on the denoised feature vector. The circumferential shear stress is analyzed by finite element analysis after inputting the denoised feature vector into the denoised feature vector. Then, the load characteristics are extracted by time-domain peak extraction and frequency-domain power spectrum analysis to finally obtain the screw load feature map.

[0052] It should be noted that step S2 includes the following steps:

[0053] The dynamic load characteristic spectrum of the screw is divided into multiple functional segments, including melt conveying functional segment, compression functional segment and metering functional segment, and the temperature field gradient and pressure field gradient correlation matrix corresponding to each functional segment is established.

[0054] Dynamic evolution simulation is performed on the correlation matrix of temperature field gradient and pressure field gradient corresponding to each functional segment, and the melt phase change process parameters corresponding to each functional segment are obtained based on the results of the dynamic evolution simulation. The melt phase change process parameters include melt phase change rate and viscosity distribution parameters.

[0055] A state-space equation is constructed, with the melt phase transformation rate and viscosity distribution parameters as input constraints. The optimal speed set is then solved by interpolating the optimal speed set to generate an intelligent speed control curve.

[0056] Specifically, functional segments are divided based on the load intensity in the screw load characteristic spectrum and combined with screw geometric parameters (such as changes in screw groove depth). For example, the region with a sharp increase in pressure gradient in the screw load characteristic spectrum is defined as the compression functional segment, the region with a steady increase in pressure is defined as the melt conveying functional segment, and the pressure peak region is defined as the metering functional segment. For the temperature field gradient and pressure field gradient collected for each segment, the correlation matrix of the temperature field gradient and pressure field gradient corresponding to each functional segment is established through multivariate regression calculation. Then, the correlation matrix of the temperature field gradient and pressure field gradient corresponding to each functional segment is dynamically evolved and simulated using finite element software to output the melt phase change rate and viscosity distribution parameters. Finally, a state space equation including screw speed, melt phase change rate, and viscosity distribution is constructed. For example, with the constraints of melt phase change completion ≥95% and viscosity fluctuation ≤±5%, the optimal speed set is solved by particle swarm optimization algorithm, and the intelligent speed control curve is generated by cubic spline interpolation.

[0057] It should be noted that step S3 includes the following steps:

[0058] The electrical parameter characteristics corresponding to the operation of the extruder drive motor winding are collected. The electrical parameter characteristics include three-phase current ripple characteristics and harmonic distortion. The current-speed correlation feature surface is constructed by combining the intelligent speed control curve.

[0059] The current-speed correlation feature surface is decomposed by fluctuation, and then the load mutation data and transient energy overshoot region are identified based on the result of the fluctuation decomposition. The load mutation data includes the load mutation point and the load mutation time, as well as the current increase at the load mutation point, the mutation duration, and the screw functional segment corresponding to the load mutation point.

[0060] An energy consumption compensation coefficient matrix is ​​established. Based on the energy consumption compensation coefficient matrix, a zero-phase difference filter bank is configured for the transient energy overshoot region. The corresponding compensation coefficient combination of the filter bank is optimized to generate an energy consumption compensation instruction set.

[0061] Specifically, in this embodiment, a high-precision current sensor is used to collect the three-phase current signal of the extruder drive motor winding. Then, based on the three-phase current signal, Fourier transform is used to extract the current ripple characteristics and calculate the harmonic distortion. At the same time, the real-time speed value of the intelligent speed control curve (e.g., 75 rpm in the metering segment and 60 rpm in the compression segment) is aligned with the electrical parameter characteristics at the corresponding time along the time axis. A current-speed correlation feature surface is constructed by three-dimensional interpolation. For example, the horizontal axis is the speed, the vertical axis is the effective current value, and the z-axis is the harmonic distortion, forming a dynamic mapping relationship between load and speed. The current-speed correlation feature surface is decomposed into fluctuations to separate high-frequency instantaneous fluctuations and low-frequency trend terms. Load mutation data is identified by setting a threshold, and the transient energy overshoot region is located by combining the energy integration method. Based on the constructed energy consumption compensation coefficient matrix, a zero-phase difference filter bank is configured for the transient energy overshoot region. The compensation coefficient combination is optimized by particle swarm optimization algorithm to generate an energy consumption compensation instruction set that includes motor torque correction and heating power adjustment values.

[0062] By adopting the above technical solution, the motor load status is captured in real time through electrical parameter characteristics, which solves the problem of limited installation of traditional mechanical sensors. The current-speed correlation feature surface intuitively presents the coupling relationship between load and speed, effectively improving the accuracy of load change point identification. Furthermore, the signal delay is eliminated by the zero phase difference filter bank, ensuring that the compensation command is synchronized with the overshoot region. The coordinated control of motor and process parameters is realized through the energy consumption compensation command set.

[0063] Furthermore, after generating the energy compensation instruction set, it also includes:

[0064] A compensation priority evaluation model is established based on load mutation data, and then the energy consumption compensation instruction set is dynamically prioritized based on the compensation priority evaluation model.

[0065] The edge computing terminal monitors and provides feedback on the compensation effect of the energy consumption compensation instruction set in real time, and then optimizes the energy consumption compensation instruction based on the monitoring and feedback results.

[0066] Specifically, a compensation priority evaluation model is constructed based on the current increase, duration, and corresponding screw functional segment of the load mutation point corresponding to the load mutation data. The inputs to the compensation priority evaluation model include mutation intensity (e.g., a 30% instantaneous current increase is considered high intensity), impact range (e.g., a mutation in the metering functional segment affects product size and has a higher weight than the feeding segment), and risk level (e.g., triggering equipment protection priority when motor harmonic distortion exceeds 5%). The weights of each indicator are calculated using the analytic hierarchy process (AHP), thereby generating a comprehensive priority score (e.g., 0-10 points, with 8 points or higher indicating emergency compensation). For example, when a load mutation lasting 2 seconds occurs in the compression functional segment... (With a 25% increase in current and a 6% increase in harmonic distortion), the compensation priority evaluation model determines the priority to be 9 points. Based on the compensation priority evaluation model, the energy consumption compensation instruction set is dynamically sorted. For example, "reduce motor torque by 15%" (emergency instruction) is placed before "fine-tune heating temperature" (regular instruction). At the same time, a real-time monitoring module is deployed on the edge computing terminal to collect the compensated electrical parameters and process parameters through current sensors and pressure transmitters, calculate the compensation effect deviation, and feed the deviation data back to the optimization module. If it is found that the compensation effect corresponding to an energy consumption compensation instruction only reaches 70% of the expected level, the instruction parameters are corrected based on the feedback results, and the priority is recalculated.

[0067] By adopting the above technical solution and setting the compensation priority evaluation model, the resource allocation problem during multiple concurrent instructions is reduced. For example, when motor overload and temperature fluctuation occur simultaneously, equipment protection instructions are executed first to avoid downtime losses. The dynamic sorting mechanism ensures that key compensation measures take effect first, and closed-loop optimization is achieved through real-time feedback from edge computing terminals, overcoming the lag of traditional open-loop compensation, thereby effectively improving equipment safety and process stability.

[0068] It should be noted that step S4 includes the following steps:

[0069] Construct a target optimization function, input the intelligent speed control curve and energy consumption compensation instruction set as constraints into the target optimization function, perform balance training on the target optimization function, and generate a dynamic energy consumption balance model.

[0070] A virtual extruder simulation environment is established through a digital twin platform. The energy consumption dynamic balance model is then subjected to transfer reinforcement learning using a set of historical process parameters. The energy consumption dynamic balance model after transfer reinforcement learning is encapsulated into a screw speed-energy consumption co-controller, and the screw speed-energy consumption co-controller is deployed to an edge computing terminal.

[0071] Furthermore, after deploying the screw speed-energy consumption co-controller to the edge computing terminal, it also includes:

[0072] Within a preset period, the screw speed-energy consumption co-controller is remotely monitored, and the corresponding optimization data of the screw speed-energy consumption co-controller is collected in real time. Based on the optimization data, the parameters of the screw speed-energy consumption co-controller are tuned.

[0073] Specifically, a target optimization function is constructed, using intelligent speed control curves (such as the upper and lower limits of speed for each functional segment) and energy consumption compensation instruction sets (such as the motor torque adjustment range) as constraint input functions. A multi-objective genetic algorithm is used to balance and train the target optimization function. Under the premise of satisfying speed-energy consumption coordinated control, an energy consumption dynamic balance model that balances efficiency and energy saving is generated. A virtual extruder simulation environment is built on a digital twin platform, mapping the screw structure, material properties, and sensor layout of the real equipment. Historical process parameter sets from the past few months are imported, and transfer reinforcement learning is performed on the energy consumption dynamic balance model. Extreme working conditions are simulated in the virtual environment, allowing the energy consumption dynamic balance model to iteratively optimize the control strategy in the simulation environment. Finally, the mature energy consumption dynamic balance model is encapsulated into a screw speed-energy consumption coordinated controller and deployed to an edge computing terminal. After deployment, within a preset period, the edge computing terminal collects optimization data such as actual energy consumption and product qualification rate in real time. Combined with the analysis results of the remote monitoring platform, the parameters of the screw speed-energy consumption coordinated controller are optimized. For example, when a certain segment of the speed control curve is found to cause large fluctuations in energy consumption, the slope of the curve is automatically corrected.

[0074] Example 2

[0075] This application also discloses a deep learning-based intelligent matching and energy consumption compensation system for extruder screw speed.

[0076] Reference Figure 2 A deep learning-based intelligent matching and energy consumption compensation system for extruder screw speed includes:

[0077] The data acquisition module is used to collect the operating status parameters and material characteristic parameters of the extruder, construct a material-equipment coupled dataset, perform screw equivalent torque analysis based on the material-equipment coupled dataset, and generate a screw load characteristic map.

[0078] The analysis module is used to simulate the evolution of the temperature-pressure coupled field based on the screw load characteristic spectrum, obtain the melt phase change process parameters, perform iterative calculation of the screw speed matching degree on the melt phase change process parameters, and generate intelligent speed control curves.

[0079] The instruction generation module is used to perform dynamic compensation analysis of energy consumption fluctuations on the intelligent speed control curve, establish an energy consumption compensation coefficient matrix, perform phase difference compensation mapping on the energy consumption compensation coefficient matrix based on the electrical parameter characteristics of the extruder drive motor winding, and generate an energy consumption compensation instruction set.

[0080] The optimization module is used to build an energy consumption dynamic balance model based on the intelligent speed control curve and energy consumption compensation instruction set, optimize the energy consumption dynamic balance model online, generate a screw speed-energy consumption co-controller, and deploy it to the edge computing terminal.

[0081] 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.

[0082] 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.

[0083] 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 method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning, characterized in that, Includes the following steps: Step S1: Collect the operating status parameters and material characteristic parameters corresponding to the extruder, construct a material-equipment coupled dataset, perform screw equivalent torque analysis based on the material-equipment coupled dataset, and generate a screw load characteristic map; Step S2: Perform temperature-pressure coupled field evolution simulation based on screw load characteristic spectrum to obtain melt phase transformation process parameters, perform screw speed matching degree iterative calculation on melt phase transformation process parameters, and generate intelligent speed control curve; Step S3: Perform dynamic compensation analysis on energy consumption fluctuations of the intelligent speed control curve, establish an energy consumption compensation coefficient matrix, perform phase difference compensation mapping on the energy consumption compensation coefficient matrix based on the electrical parameter characteristics of the extruder drive motor winding, and generate an energy consumption compensation instruction set. Step S4: Construct an energy consumption dynamic balance model based on the intelligent speed control curve and energy consumption compensation instruction set, optimize the energy consumption dynamic balance model online, generate a screw speed-energy consumption co-controller, and deploy it to the edge computing terminal.

2. The method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning according to claim 1, characterized in that, Step S1 includes the following steps: The axial distributed pressure value corresponding to the extruder screw is collected by a pressure sensor array, and the melt index and thermal conductivity parameters of the material are obtained by an infrared spectroscopy sensor. Then, a material-equipment coupled dataset is constructed based on the axial distributed pressure value, the melt index and thermal conductivity parameters of the material. The material-equipment coupled dataset is denoised, and a denoised feature vector is generated based on the denoising result. An equivalent mechanical model is constructed, and the denoised feature vector is input into the equivalent mechanical model for circumferential shear stress analysis. Load features are extracted from the circumferential shear stress analysis results to generate a screw load feature map.

3. The method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning according to claim 1, characterized in that, Step S2 includes the following steps: The dynamic load characteristic spectrum of the screw is divided into multiple functional segments, including melt conveying functional segment, compression functional segment and metering functional segment, and the temperature field gradient and pressure field gradient correlation matrix corresponding to each functional segment is established. Dynamic evolution simulation is performed on the correlation matrix of temperature field gradient and pressure field gradient corresponding to each functional segment, and the melt phase change process parameters corresponding to each functional segment are obtained based on the results of the dynamic evolution simulation. The melt phase change process parameters include melt phase change rate and viscosity distribution parameters. A state-space equation is constructed, with the melt phase transformation rate and viscosity distribution parameters as input constraints. The optimal speed set is then solved by interpolating the optimal speed set to generate an intelligent speed control curve.

4. The method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning according to claim 1, characterized in that, Step S3 includes the following steps: The electrical parameter characteristics corresponding to the operation of the extruder drive motor winding are collected. The electrical parameter characteristics include three-phase current ripple characteristics and harmonic distortion. The current-speed correlation feature surface is constructed by combining the intelligent speed control curve. The current-speed correlation feature surface is subjected to fluctuation decomposition, and then the load mutation data and transient energy overshoot region are identified based on the results of the fluctuation decomposition. The load mutation data includes the load mutation point and the load mutation time. An energy consumption compensation coefficient matrix is ​​established. Based on the energy consumption compensation coefficient matrix, a zero-phase difference filter bank is configured for the transient energy overshoot region. The corresponding compensation coefficient combination of the filter bank is optimized to generate an energy consumption compensation instruction set.

5. The method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning according to claim 4, characterized in that, After generating the energy consumption compensation instruction set, it also includes: A compensation priority evaluation model is established based on load mutation data, and then the energy consumption compensation instruction set is dynamically prioritized based on the compensation priority evaluation model. The edge computing terminal monitors and provides feedback on the compensation effect of the energy consumption compensation instruction set in real time, and then optimizes the energy consumption compensation instruction based on the monitoring and feedback results.

6. The method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning according to claim 1, characterized in that, Step S4 includes the following steps: Construct a target optimization function, input the intelligent speed control curve and energy consumption compensation instruction set as constraints into the target optimization function, perform balance training on the target optimization function, and generate a dynamic energy consumption balance model. A virtual extruder simulation environment is established through a digital twin platform. The energy consumption dynamic balance model is then subjected to transfer reinforcement learning using a set of historical process parameters. The energy consumption dynamic balance model after transfer reinforcement learning is encapsulated into a screw speed-energy consumption co-controller, and the screw speed-energy consumption co-controller is deployed to an edge computing terminal.

7. The method for intelligent matching of extruder screw speed and energy consumption compensation based on deep learning according to claim 6, characterized in that, After deploying the screw speed-energy consumption co-controller to the edge computing terminal, it also includes: Within a preset period, the screw speed-energy consumption co-controller is remotely monitored, and the corresponding optimization data of the screw speed-energy consumption co-controller is collected in real time. Based on the optimization data, the parameters of the screw speed-energy consumption co-controller are tuned.

8. A deep learning-based intelligent matching and energy consumption compensation system for extruder screw speed, applied to the deep learning-based intelligent matching and energy consumption compensation method for extruder screw speed as described in any one of claims 1-7, characterized in that, include: The data acquisition module is used to collect the operating status parameters and material characteristic parameters of the extruder, construct a material-equipment coupled dataset, perform screw equivalent torque analysis based on the material-equipment coupled dataset, and generate a screw load characteristic map. The analysis module is used to simulate the evolution of the temperature-pressure coupled field based on the screw load characteristic spectrum, obtain the melt phase change process parameters, perform iterative calculation of the screw speed matching degree on the melt phase change process parameters, and generate intelligent speed control curves. The instruction generation module is used to perform dynamic compensation analysis of energy consumption fluctuations on the intelligent speed control curve, establish an energy consumption compensation coefficient matrix, perform phase difference compensation mapping on the energy consumption compensation coefficient matrix based on the electrical parameter characteristics of the extruder drive motor winding, and generate an energy consumption compensation instruction set. The optimization module is used to build an energy consumption dynamic balance model based on the intelligent speed control curve and energy consumption compensation instruction set, optimize the energy consumption dynamic balance model online, generate a screw speed-energy consumption co-controller, and deploy it to the edge computing terminal.

9. A computer-readable storage medium, characterized in that: The device stores instructions that, when executed on a computer, cause the computer to perform the deep learning-based intelligent matching and energy consumption compensation method for extruder screw speed as described in any one of claims 1 to 7.

Images