Visualization Of Abnormal Flame Signals Based On Euler Video Amplification Technology

Under the background of energy shortage and environmental issue,the safty and efficiency of energy utilization and the reduction of pollution emission have become one of the hot topics today.With the technology developing,the complexity of industrial combustion systems has improved dramatically,which requires higher safty and efficiency for the combustion process.The combustion process is a continuous process that changes with time.Therefore,the change of condition is inevitably a continuous process,not an unpredictable "absolute abrupt change".However,the signal causing the condition change is usually weak at the beginning,which is difficult to recognize and monitoring.As one of the important parameters for the combustion process,the flame pulsation frequecny can be used to diagnose the combustion stability.Therefore,it is important to identify the weak and unusual flame freqeucny signals during the combutaion process for combustor maintenance arragement and safety accident prevention.At the beginning,this paper discusses the fundamental knowedge of combustion and the importance of combustion flame stability,then the developments and shortcomings of flame pulsation frequency detection methods are also analysed.According to this,the weak and unusual flame frequency signal detection and visualization method is proposed based on the Euler video amplification technology(EVM).Firstly,the Digital Flame Color Distribution model(DFCD)is used to separate and extract the red-yellow flame region.Then,the Eulerian video amplification technique(EVM)is applied to vedio siganl spectrum of this extracted flame region for visualization.After segentation and amlification,the video average gray intensities of this flame region are transferred by the Fast Fourier Transform(FFT)to obtain the amplified flame pulsation frequency specturm.The combustion condition is then diagnosed by comparing the frequency components.After that,the differential image processing is applied to the flame images that include unusual flame frequency siganls.At last,the accurate position of the unusual signal in the flame region will be detected by the morphological processing method.Through the theoretical research and experimental verification,the results show that the proposed method based on the DFCD and EVM techniques can identify and amplify the weak signals in the flame pulsation region effectively,and it also can realize the visualization of this region clearly.Therefore,this proposed method provides a new idea and method for detecting the weak and unusual flame frequency signals,and it is also an potential method that can greatly reduce the hardware cost of combustion measurement equipments.