| Continuous casting technology is one of the fastest growing areas of steelmakingtechnology, in the process of continuous casting, breakout is a major productionaccidents. It will not only reduce the production efficiency, but also seriously affectthe quality of the casting. Research on breakout prediction system is of greateconomic benefit and social benefit.The paper firstly introduces the basic knowledge of casting, then analysis of theformation mechanism of sticking breakout. In contrast to commonly used in industrialthermocouple, select the K type thermocouple as the temperature measurement devicesystem, which temperature response slope curve is smoother. Analysis the model ofthermocouples embedded in the crystallizer in detail. Analysis and Research on thebreakout prediction system hardware, algorithms and software design.In the temperature acquisition unit, the latest dedicated digital conversion chipMAX31855K was selected, it not only solved the difficulties of temperaturemeasurement, but also simplifies the circuit design. The paper presents thetemperature acquisition unit programming flowcharts, and use extreme averagefiltering algorithm to eliminate noise interference. In the design of TCP/IPcommunication unit, dual port RAM CY7C136was used as data sharing memory, andpresented the of CY7C136peripheral circuit design. CP2200is chosen as the Ethernetcontroller chip, CP2200peripheral circuit design and flowchart of software designwere given.In the software design process, the traditional breakout prediction methods ofsteel led to the breakout prediction system based on BP neural network. Afteranalyzing the defects of BP algorithm, improved particle swarm was proposed tooptimize BP neural network, and the performance test of the improved network wasconducted. In the design of client, the main interface, the system software of themeasured temperature curve, bonding temperature curve and false alarm temperaturecurves are given, and make the corresponding explanation. |
PDF Full Text Request