Fusion-controlled Study, Based On The Embedded System Furnace

Steel is the foundational raw material of the national economy. The production process of the reheating furnace is the core part of the cement produce.It is a complex engineering thermodynamics reaction process with great inertia, pure lag, non-linearity and strong coupling. It is difficult to establish accurate and complex model, so traditional control ways can hardly control it accurately. In this paper, we study the reheating furnace on data fusion control base on neural network deeply.Based on the desing of automatic contorl system of reheating furnace,thesis firstly introduces the craftwork of reheating furnace and the contorl requests of computer-aided control system. Secondly, the actual conortl system srtategies are inortduced in detail. In this paper, we use neural network it has abilities of learn memory, self-organizing and self-adaptation base on the knowledge of neural network. Through comprehensive judge and global analysis to form a new control strategy. We fuse the temperature of reheating furnace, pressure of reheating furnace,which make the key role in the working process of reheating furnace, it solves the problem of choice, the control ways inaccuracy, which hardly reacts whole engineering status by single sensor.Through classify data by ART-2 and BP neural network to get the accurate engineering status in the location, and training them. The collected data into space of input information-space of classification and space of cluster then dealing with them in whole repository, forming control strategy tactics in detail, composing complete control system.The hardware platform UP-Net ARM 2410-S is used as the target and the PC installed REDHAT 9 is used as the host. The cross-compile tool is installed on the host to establish the cross-compiling environment. It realizes the hardware design of reheating furnace based on platform UP-Net ARM 2410-S, the development of man-machine interface with Qt as the development tool.