Search

CN-121979147-A - Intelligent fault diagnosis and repair system for full-biodegradable composite material production line

CN121979147ACN 121979147 ACN121979147 ACN 121979147ACN-121979147-A

Abstract

The invention relates to the technical field of intelligent control of polymer material processing, and discloses an intelligent fault diagnosis and repair system of a full-biodegradable composite material production line. The method comprises the steps of constructing a unified data lake with synchronous time by a data acquisition module at an edge layer, generating a dynamic health reference curve by fusing a non-isothermal non-Newtonian fluid transportation mechanism and a deep learning model by a digital twin engine, calculating a residual sequence of actual monitoring data and the reference curve by a fault diagnosis module, identifying a fault mode by multi-mode feature fusion, generating a control compensation instruction by a self-repairing decision module based on the fault mode, and issuing a closed-loop control by safety verification. The invention solves the diagnosis problem caused by material characteristic fluctuation through a mechanism and data driving fusion technology, realizes early fault early warning, process parameter self-adaptive compensation and predictive maintenance in the production process, and improves the consistency of product quality and production stability.

Inventors

  • ZHU FU
  • ZHANG CHANGPING

Assignees

  • 首掌科技创新(济南)有限公司

Dates

Publication Date
20260505
Application Date
20260203

Claims (10)

  1. 1. An intelligent fault diagnosis and repair system of a full-biodegradable composite material production line is characterized by comprising a physical production line layer (10), an edge calculation and intelligent control layer (20) and a cloud platform data center layer (30); The physical production line layer (10) comprises a reaction extrusion unit, and the physical production line layer (10) acquires equipment state time sequence data, process parameter time sequence data and on-line quality time sequence data reflecting the reaction extrusion process of the full-biodegradable material through a sensor network; The edge computation and intelligent control layer (20) comprises: The data acquisition and preprocessing module (21) is used for receiving the equipment state time sequence data, the process parameter time sequence data and the online quality time sequence data to construct a time-synchronous unified data lake; a digital twin engine (22) for generating a health state reference curve based on the unified data lake using a mechanism and data driven hybrid model based on a non-isothermal non-newtonian fluid transport principle; The fault diagnosis and health management module (23) is used for outputting a fault mode by utilizing the actual monitoring data in the unified data lake and the residual sequence of the health state reference curve; and the process parameter self-repairing decision module (24) is used for generating a control compensation instruction according to the fault mode and issuing the control compensation instruction to the physical production line layer (10) to execute closed-loop control.
  2. 2. The intelligent fault diagnosis and repair system of a full-biodegradable composite material production line according to claim 1, wherein the process of the data acquisition and preprocessing module (21) receiving the equipment state time sequence data, the process parameter time sequence data and the on-line quality time sequence data to construct a time-synchronized unified data lake comprises: marking the equipment state time sequence data, the process parameter time sequence data and the online quality time sequence data from the physical production line layer (10) by utilizing a unified clock source, performing microsecond time stamp alignment operation on data sources with different sampling rates by utilizing an interpolation algorithm, and constructing heterogeneous multi-source data aggregation into the unified data lake which is strictly synchronous in a time dimension.
  3. 3. The intelligent fault diagnosis and repair system of a fully biodegradable composite material production line according to claim 1, wherein the step of operating a mechanism model path in the process of generating a health state reference curve based on the unified data lake by the digital twin engine (22) using a mechanism based on a non-isothermal non-newton fluid transport principle and a data-driven hybrid model comprises: establishing a one-dimensional distributed parameter model in the screw extrusion process, and discretizing the screw into a plurality of finite control bodies along the axial direction; According to the screw rotating speed and the temperature feedback value of the temperature zone acquired in real time, calculating theoretical viscosity distribution along the axial direction of the screw by using a rheological state equation describing the functional relation between the apparent viscosity of the material, the shear rate and the temperature; and outputting a theoretical melt pressure curve and a theoretical torque curve under the current working condition as physical references of the health state reference curve through pressure gradient integration and viscous dissipation integration along the flow channel.
  4. 4. An intelligent fault diagnosis and repair system for a fully biodegradable composite material production line according to claim 3, wherein the step of operating the data driven path in the process of generating the health status reference curve based on the unified data lake by the digital twin engine (22) using a mechanism based on a non-isothermal non-newton fluid transport principle and a data driven hybrid model comprises: Adopting a long-short-time memory network enhanced based on an attention mechanism to receive the equipment state time sequence data and the process parameter time sequence data in the unified data lake as input vectors; And capturing long-term dependency and dynamic hysteresis effect in the input vector by using a long-short-term memory network unit, identifying nonlinear influence weight of the key process parameter on the quality index by using an attention mechanism module, and outputting a predicted value of the key quality parameter to correct the physical reference.
  5. 5. The intelligent fault diagnosis and repair system of a full-biodegradable composite material production line according to claim 1, wherein the step of calculating the residual sequence in the process of outputting a fault pattern by the fault diagnosis and health management module (23) using the residual sequence of the actual monitoring data and the health status reference curve in the unified data lake comprises: -receiving said actual monitoring data from said data acquisition and preprocessing module (21) and data from said state of health reference curve of said digital twin engine (22); And calculating the difference value between the actual value and the theoretical health reference value according to the key variables of the host torque, the inlet pressure and the outlet pressure of the melt pump, and generating the residual sequence reflecting the equipment performance degradation or abnormal disturbance by stripping parameter fluctuation caused by normal process setting adjustment.
  6. 6. The intelligent fault diagnosis and repair system of a full-biodegradable composite material production line according to claim 5, characterized in that the step of extracting features from the residual sequence during the fault pattern output by the fault diagnosis and health management module (23) using the residual sequence of the actual monitoring data in the unified data lake and the health state reference curve comprises: processing is performed in parallel through three channels: the first channel performs windowing on the residual sequence to extract a mean value, a variance, a skewness, a kurtosis and a trend slope as time domain statistical characteristics; The second channel performs fast Fourier transform on the periodic fluctuation signal to extract the energy duty ratio of the preset frequency band as the frequency domain energy characteristic; And the third channel adopts a dynamic time warping algorithm to calculate the similarity distance between the actually collected spectrum curve and the spectrum curve of the standard qualified product as the track abnormal characteristic.
  7. 7. The intelligent fault diagnosis and repair system of a full-biodegradable composite material production line according to claim 6, wherein the step of outputting the fault pattern based on the extracted features in outputting the fault pattern by the fault diagnosis and health management module (23) using the residual sequence of the actual monitoring data in the unified data lake and the health status reference curve comprises: fusing the time domain statistical features, the frequency domain energy features and the track abnormal features to form a multi-mode feature vector; inputting the multi-modal feature vectors into a pre-trained gradient lifting tree classifier, mapping the multi-modal feature vectors into a predefined fault mode space by the gradient lifting tree classifier, and outputting confidence probability of each fault mode; and setting a probability threshold, and judging that the fault mode occurs when the prediction probability of the fault mode of a certain type exceeds the probability threshold.
  8. 8. The intelligent fault diagnosis and repair system of a fully biodegradable composite material production line according to claim 1, wherein the step of executing deterministic repair based on a rule base in the process of generating a control compensation instruction by the process parameter self-repair decision module (24) according to the fault mode comprises: When the fault mode is that a filter screen is blocked and the confidence coefficient probability exceeds a preset threshold value, automatically generating the control compensation instruction to control the variable frequency motor of the melt gear pump to increase the rotating speed, and maintaining the die head pressure within a set value range; and calculating the residual service life of the filter screen according to the current pressure difference rising rate and the pump speed rising amplitude.
  9. 9. The intelligent fault diagnosis and repair system of a fully biodegradable composite material production line according to claim 1, wherein the step of performing a reinforcement learning-based compensation decision in the process of generating a control compensation instruction by the process parameter self-repair decision module (24) according to the fault mode comprises: Aiming at the diagnosed quality deviation caused by the fluctuation of the intrinsic characteristics of the raw materials, a depth deterministic strategy gradient network is called as a reinforcement learning compensator; And receiving the complete process state vector and the mass deviation value at the current moment as input states, and outputting motion vectors containing multidimensional adjustment amounts, wherein the motion vectors correspond to the adjustment amounts of the rotating speed of a host screw and the set temperature of a reaction zone, and compensate rheological property changes caused by material characteristics.
  10. 10. The intelligent fault diagnosis and repair system of a fully biodegradable composite material production line according to claim 1, wherein the step of executing a safety interaction mechanism in the process of issuing the control compensation instruction to the physical production line layer (10) by the process parameter self-repair decision module (24) to perform closed-loop control comprises: Before issuing the control compensation instruction, the edge calculation and intelligent control layer (20) verifies whether the control compensation instruction is within a preset process safety envelope curve range; And popping up a highlight prompt on an operation interface, starting a countdown waiting confirmation period, and automatically issuing and executing the control compensation instruction if the countdown is finished and the overrule operation of an operator is not received.

Description

Intelligent fault diagnosis and repair system for full-biodegradable composite material production line Technical Field The invention relates to the technical field of intelligent control of processing of high polymer materials, in particular to an intelligent fault diagnosis and repair system of a full-biodegradable composite material production line. Background At present, the preparation of the fully biodegradable multiphase composite materials such as PBAT, PLA and the like mainly adopts a reaction extrusion process. The production process integrates multi-component feeding, melt plasticizing, reaction grafting and other units, and relates to thermal, force and rheological multi-field coupling. The process window is usually narrow, and the physical properties of raw materials in different batches are different, so that the degradation performance and mechanical strength of the final product are directly determined by the stability of the production process. For the above production process, the existing control system is mostly operated based on the PLC architecture. The system collects temperature, pressure, host current and other data in real time through the sensor, and utilizes a PID algorithm to carry out feedback adjustment on the temperature of the heating zone and the rotating speed of the screw. The fault monitoring mainly depends on preset upper and lower limit thresholds, and when the monitored value exceeds a set range, an alarm or shutdown interlocking is triggered. And an operator monitors the running state according to the data curve displayed by the interface, and manually intervenes and adjusts the technological parameters according to the off-line detection result. However, the prior art has shortcomings in coping with complex conditions. The fixed threshold monitoring method is insensitive to gradual change faults, and is difficult to give out a warning at the early stage of filter screen blockage or at the stage of screw slight abrasion, and the stopping accident is often delayed. Due to the lack of sensing and dynamic compensation capability for intrinsic characteristic fluctuation of raw materials, the influence of raw material viscosity or water content change on the melt index of the product cannot be eliminated by the traditional PID control, so that the quality consistency among batches is poor. In addition, the multivariable coupling makes the tracing of the fault source difficult, the investigation efficiency is low by relying on manual experience alone, and the operation and maintenance cost is high due to the lack of a scientific life prediction mechanism, so that the active predictive maintenance cannot be realized. Therefore, the invention provides an intelligent fault diagnosis and repair system of a full-biodegradable composite material production line, which aims to solve the defects in the prior art. Disclosure of Invention Aiming at the defects of the prior art, the invention provides an intelligent fault diagnosis and repair system of a full-biodegradable composite material production line, which solves the problems of fault early warning hysteresis, poor product quality consistency caused by raw material fluctuation and equipment performance degradation and high passive operation and maintenance cost in the existing full-biodegradable composite material production. In order to achieve the purpose, the intelligent fault diagnosis and repair system of the full-biodegradable composite material production line comprises a physical production line layer, an edge calculation and intelligent control layer and a cloud platform data center layer; The physical production line layer comprises a reaction extrusion unit, and acquires equipment state time sequence data, process parameter time sequence data and online quality time sequence data reflecting the reaction extrusion process of the fully biodegradable material through a sensor network; the edge calculation and intelligent control layer comprises: The data acquisition and preprocessing module is used for receiving the equipment state time sequence data, the process parameter time sequence data and the online quality time sequence data to construct a time-synchronous unified data lake; The digital twin engine is used for generating a health state reference curve based on the unified data lake by utilizing a mechanism and data driving mixed model based on a non-isothermal non-Newtonian fluid transportation principle; The fault diagnosis and health management module is used for outputting a fault mode by utilizing the actual monitoring data in the unified data lake and the residual sequence of the health state reference curve; And the process parameter self-repairing decision module is used for generating a control compensation instruction according to the fault mode and issuing the control compensation instruction to the physical production line layer to execute closed-loop control. By adopting the technical scheme, as a