Research On Intelligent Detection Method Based On Flame Image Recognition For Ceramic Roller Kiln Burning Zone

The Ceramic roller kiln is widely used in the production of ceramic products.The rotating rolls are used as carrying tools of ceramic products in it. In the process of production, ceramic products are placed on the top of intensive flame resistant roller. The rotation rollerstransfer ceramic products from the preheatingzone of kiln head to cooling zone of kiln tail so as to make ceramic productsfired.Ceramic products quality mainly depend on sintering conditions of ceramic productsin ceramic roller kiln, and sintering conditions of ceramic product are closely related to firing zone working condition, hence,the detection of firing zone working condition is a key link in ceramic kiln production process. Therefore, the researchon the effective recognition method for roller kiln burning zone is with great important significance. In this paper, for the defects of ceramic kiln temperature detection, it is proposed that the flame image recognition technology of simulating artificial-look-fire is fused with multi-point temperature detection method to recognize the working conditions of ceramic kiln firing zone. At the same time,the intelligent detection system based on multi-sensor is designed to recognize firing zone condition recognition of ceramic roller kiln, and the simulation platform is also set up. Main results are as follows.1. The key process data fusion method is proposed for firing zone temperature of ceramic roller kiln based on the multi-sensor. The multipoint temperatures of firing zone are collected by thermocouple sensors, and temperature data collected are fused by information fusion algorithm so as to get the key process data.2. According to the flame image characteristics of ceramic roller kiln, At first, flame image is denoised by inter-frame smooth and median filter method. Then, the image segmentation is completed by K-means clustering algorithm. At last, characteristic value of flame part and r, g, b are extracted.3. The information fusion fuzzy expert system has been built. By the system, the flame image characteristics and the key process data are fused to get the effective information to recognize the combustion zone working condition. Through Matlab fuzzy toolbox simulation, the simulation results show the feasibility and effectiveness of the method proposed.