Search

CN-121979153-A - Flexible intelligent manufacturing system based on AI intelligent control

CN121979153ACN 121979153 ACN121979153 ACN 121979153ACN-121979153-A

Abstract

The invention discloses a flexible intelligent manufacturing system based on AI intelligent control, which belongs to the technical field of industrial intelligent manufacturing, wherein the system realizes full-flow sensing, transmission, analysis and control closed loop of manufacturing through an industrial Internet of things framework, breaks through the traditional flexible manufacturing bottleneck by combining an AI algorithm, and comprises an industrial Internet of things sensing layer, an edge transmission layer, an AI intelligent control layer, a flexible execution layer and a cloud cooperation layer, wherein the sensing layer uses multiple types of sensors to collect data in a global and precise manner, the edge transmission layer is based on a specific protocol and a 5G gateway to realize low-delay and high-reliability transmission and preprocessing, the AI intelligent control layer analyzes data in real time to perform fault prediction and the like, the flexible execution layer responds to instructions to realize flexible cooperation, and the cloud cooperation layer gathers data and an iterative model. The invention solves the technical problems of traditional flexible manufacturing, improves the flexibility, the intelligent level and the production efficiency of the system, reduces the cost and is suitable for flexible production scenes in multiple fields.

Inventors

  • WEI MENGQI

Assignees

  • 唐山祺德科技有限公司

Dates

Publication Date
20260505
Application Date
20260206

Claims (10)

  1. 1. A flexible intelligent manufacturing system based on intelligent AI control is characterized by taking global sensing and intelligent AI control cooperation of an industrial Internet of things as cores to realize flexible and intelligent closed-loop control of a whole manufacturing process, the system comprises an industrial Internet of things sensing layer, an edge transmission layer, an intelligent AI control layer, a flexible execution layer and a cloud cooperation layer in two-way communication with the edge transmission layer and the intelligent AI control layer, which are sequentially connected in a communication mode, wherein the industrial Internet of things sensing layer is used for collecting various data of the whole manufacturing process, the edge transmission layer is used for low-delay transmission and local preprocessing of the data, the intelligent AI control layer is used for realizing intelligent decision and control instruction generation based on the Internet of things data, the flexible execution layer is used for responding to control instructions to realize flexible production, and the cloud cooperation layer is used for data aggregation, model iteration and scene crossing cooperation.
  2. 2. The flexible intelligent manufacturing system based on AI intelligent control of claim 1, wherein the industrial Internet of things sensing layer comprises an equipment state sensing module, a material tracing sensing module, an environment parameter sensing module, a process parameter sensing module and a safety state sensing module, wherein the equipment state sensing module adopts a vibration sensor, a current sensor and a temperature sensor, is connected with key parts of production equipment through an Internet of things interface and is used for collecting parameters such as equipment rotating speed, vibration frequency, working current, surface temperature and operation time, the material tracing sensing module adopts an RFID reader and an RFID tag, the RFID tag is attached to materials, semi-finished products and finished products, information such as material model, specification, position, circulation time and processing procedure is collected, the environment parameter sensing module adopts a temperature and humidity sensor, a dust sensor and a gas pressure sensor and is used for collecting environment parameters such as temperature and humidity, dust concentration and gas pressure of a production workshop, the process parameter sensing module adopts a pressure sensor, a flow sensor and a displacement sensor and is used for collecting real-time process parameters such as cutting pressure, flow and processing displacement, and the safety state sensing module adopts an infrared sensor, a smoke sensor, a position sensor and fire hazard safety protection personnel and a workshop safety protection device.
  3. 3. The flexible intelligent manufacturing system based on AI intelligent control of claim 1, wherein the edge transmission layer comprises a 5G edge gateway, an industrial Ethernet switch, an edge preprocessing unit and a communication protocol conversion module, wherein the 5G edge gateway adopts an industrial grade gateway supporting slicing technology to realize local access and edge unloading of sensing layer data, the transmission delay is controlled within 10ms, the communication protocol conversion module supports bidirectional conversion of multiple industrial communication protocols of OPC UAover TSN and Modbus, profinet to solve the problem of communication compatibility between different brands of Internet of things equipment and production equipment, the edge preprocessing unit adopts an embedded chip to perform denoising, normalization and outlier rejection processing on original data acquired by the Internet of things, extracts key characteristic data and then transmits the key characteristic data to the AI intelligent control layer, and meanwhile encrypts and uploads the original data to a cloud intelligent control layer.
  4. 4. The flexible intelligent manufacturing system based on AI intelligent control of claim 1 is characterized in that the AI intelligent control layer comprises a data fusion module, an AI prediction control module, a dynamic scheduling module, a process optimization module and an instruction generation module, wherein the data fusion module is used for carrying out fusion processing on multi-source heterogeneous data of the Internet of things transmitted by an edge transmission layer by adopting a weighted fusion algorithm to generate a unified-format characteristic data set, the AI prediction control module is used for analyzing equipment state data and process parameter data based on an LSTM deep learning algorithm to realize equipment failure prediction and process deviation prediction, the prediction accuracy is not lower than 98%, the dynamic scheduling module is used for generating an optimal production scheduling scheme in real time based on reinforcement learning algorithm and combining order information, material states and equipment load data acquired by the Internet of things to support dynamic switching of multi-variety mixed line production, the process optimization module is used for dynamically optimizing a process parameter threshold value based on the deep learning algorithm by analyzing the association relation between historical process parameters and product quality data, and the instruction generation module is used for converting an AI result into a standardized control instruction and sending the standardized control instruction to a flexible decision-making layer.
  5. 5. The flexible intelligent manufacturing system based on the AI intelligent control is characterized in that the flexible execution layer comprises a modularized production unit, a flexible conveying module, an intelligent tool fixture, an AGV logistics robot and an execution control module, wherein the modularized production unit is designed by adopting a standardized interface, each module is provided with an independent communication interface and a control interface of the Internet of things, the rapid recombination is supported through AI control instructions, the recombination time is not more than 48 hours, the flexible conveying module is a telescopic conveying line and is cooperated with the AGV logistics robot to realize flexible material transfer, the intelligent tool fixture is of an electric adjustment structure and is used for receiving AI control instructions through the Internet of things to realize rapid clamping of workpieces with different specifications, the execution control module is communicated with the AI intelligent control layer, receives the control instructions and drives each execution part to cooperatively work, and simultaneously collects real-time data in the execution process and feeds back to an edge transmission layer.
  6. 6. The flexible intelligent manufacturing system based on AI intelligent control according to claim 1, wherein the cloud cooperation layer comprises a data storage module, a model iteration module, a cross-factory cooperation module and a visual monitoring module, the data storage module adopts a distributed database to store original data acquired by an industrial Internet of things sensing layer, feature data after edge preprocessing, AI control decision data and production process data, storage delay is not more than 1s, the model iteration module continuously trains and iteratively updates a prediction model, a scheduling model and an optimization model of the AI intelligent control layer based on mass data converged by the cloud, an iteration period can be dynamically adjusted according to production requirements, the cross-factory cooperation module realizes data sharing and instruction cooperation of multiple factories through an industrial Internet of things and supports flexible production scheduling of the cross-factory, and the visual monitoring module displays information such as the Internet of things sensing data, equipment running states, AI decision results and production progress to users through a Web end and a mobile end and supports abnormal alarm and remote control.
  7. 7. The flexible intelligent manufacturing system based on AI intelligent control of claim 2, wherein all sensors of the industrial Internet of things sensing layer are designed with low power consumption, a battery power supply and wire power supply dual mode is supported, the sensors and an edge transmission layer are communicated by adopting an encryption transmission protocol, data leakage is prevented, the RFID tag adopts a high-frequency passive tag, the reading distance is 0.5-5m, simultaneous reading of multiple tags is supported, the reading speed is not lower than 100 pieces/min, the sampling frequency of the sensors of the process parameter sensing module can be dynamically adjusted through the AI intelligent control layer, the sampling frequency range is 10-100Hz, and the monitoring requirements of different processes are met.
  8. 8. The flexible intelligent manufacturing system based on AI intelligent control as set forth in claim 3, wherein the edge transmission layer further comprises a data caching unit, the data caching unit can cache the sensing data of the Internet of things for at least 72 hours when the 5G network is interrupted, the network is automatically synchronized to the AI intelligent control layer and the cloud cooperation layer after recovery, the industrial Ethernet switch adopts a gigabit industrial grade switch to support redundancy backup, ensure the reliability of data transmission and avoid data loss caused by single-point faults.
  9. 9. The flexible intelligent manufacturing system based on the AI intelligent control system as set forth in claim 4, wherein the AI intelligent control layer further comprises a fault emergency processing module, when the AI prediction control module detects that equipment fault hidden danger or data of the Internet of things are abnormal, the fault emergency processing module immediately generates an emergency control instruction, the emergency control instruction is issued to the flexible execution layer, shutdown protection of fault equipment and automatic starting of standby equipment are achieved, alarm information is sent to related management staff through the cloud coordination layer, and emergency response time is not more than 500ms.
  10. 10. The flexible intelligent manufacturing system based on AI intelligent control according to any one of claims 1-9, wherein the system further comprises a security protection layer, the security protection layer comprises a data encryption module, an access control module and an intrusion detection module, the data encryption module encrypts data collected by an industrial Internet of things sensing layer, data transmitted by an edge transmission layer and data stored by a cloud end to end, the access control module adopts a hierarchical authorization mechanism to strictly control system access rights of different users, and the intrusion detection module monitors Internet of things transmission links and system ports in real time, timely discovers and blocks illegal intrusion behaviors and ensures safe and stable operation of the system.

Description

Flexible intelligent manufacturing system based on AI intelligent control Technical Field The invention belongs to the technical field of industrial intelligent manufacturing, and particularly relates to a flexible intelligent manufacturing system based on intelligent AI control. Background Along with the transformation of manufacturing industry to multi-variety, small batch and customization, flexible intelligent manufacturing becomes a core direction breaking through the bottleneck of traditional rigid production, and industrial Internet of things is used as a neural center of flexible manufacturing to bear the key effects of sensing and transmitting production full-flow data, AI intelligent control provides core support for dynamic adjustment and intelligent decision of flexible production, the existing flexible intelligent manufacturing system preliminarily fuses the industrial Internet of things and AI technology, but has a plurality of technical defects, and is difficult to meet the production requirements of high efficiency, accuracy and flexibility, and the method comprises the following steps: the perception of the industrial Internet of things is incomplete and the precision is insufficient: The existing system is mainly used for simply sensing production equipment, lacks of global coverage of material circulation, production environment, technological parameters and safety states, is single in sensor type and poor in communication compatibility, cannot realize data intercommunication between different brands of equipment and sensors, causes collected data fragmentation, cannot provide complete and accurate data support for AI control, and meanwhile, partial sensors cannot be fixed in sampling frequency and cannot be dynamically adjusted according to changes of production technology, so that instantaneous technological deviation and equipment abnormality are difficult to capture. 2. Data transmission lags, poor reliability: The conventional system mostly adopts the traditional industrial Ethernet to transmit data, has higher transmission delay (usually more than 50 ms), cannot meet the real-time requirements of equipment recombination and process adjustment in flexible production, lacks the capability of edge preprocessing, and directly uploads massive raw data to the cloud, so that the network bandwidth pressure is increased, data loss and transmission congestion are easily caused, meanwhile, the communication protocol is not uniform, and the data interaction between different modules has barriers, so that the transmission efficiency is further reduced. 3. AI and industry thing networking are in coordination with deficiency: In the existing system, AI control and industrial Internet of things are designed in a separated mode, an AI algorithm only carries out decision making based on a small amount of preprocessed data, dynamic data acquired by the Internet of things cannot be received in real time and a model is iterated quickly, so that AI decision making is disjointed from an actual production scene, flexible adjustment is lagged, meanwhile, an AI decision making result cannot be fed back to an Internet of things perception layer quickly, closed loop cooperation of perception-analysis-decision-control cannot be achieved, and a cooperative effect of the AI decision making and the industrial Internet of things cannot be exerted. 4. Flexible execution is not tight with AI control linkage: The flexible execution unit of the existing system is controlled by a fixed program, dynamic scheduling and process optimization instructions generated by AI cannot be responded quickly, the modularization recombination efficiency of equipment is low (usually more than 72 hours are needed), the switching loss of multi-variety mixed line production is large, real-time data in the execution process cannot be fed back to an AI control layer in time through the Internet of things, so that the control accuracy is insufficient, and the consistency of product quality is poor. 5. The system has the defects of safety and expansibility: The existing system lacks of full-flow safety protection for industrial Internet of things data, is easy to leak, tamper and the like in the data transmission and storage processes, is fixed in system architecture, cannot be flexibly expanded according to the requirements of different industries and different production scenes, and is poor in adaptability. In addition, the prior art is like an 'AI intelligent control-based flexible intelligent manufacturing system' with a publication number of CN121187256A, only focuses on the prediction and adjustment of the position offset of equipment, does not realize the global perception and the depth cooperation of AI of the industrial Internet of things, and the 'real-time analysis and issuing method and system of process parameters in the automobile flexible manufacturing' with a publication number of CN121411356A only aims at the process parameter