Research On Parameters Detectionalgorithm Of Gas-solid Two-phase Flow

Gas-solid two-phase flow system has been widespread in nature and industrialproduction process. With the rapid development of science and technology, thedetection of gas-solid flow parameters, especially the detection of gas-solid two-phaseflow in the industrial production process, is becoming increasingly important toimprovement of productivity, reducing energy consumption and environmentalmonitoring. However, the complex characteristics of gas-solid flow make detectionmore difficult. Numbers of research on gas-solid flow detection have been done, butmostly research is operated in a laboratory environment, which can’t fully satisfy theneed of industry scene. In practical applications, the industrial environment iscomplex and varieties of outside factors have an impact of the detection of gas-solidflow, with the impact of gas-solid flow fluid itself, parameter detection for gas-solidflow corresponding becomes more difficult. At that case, the traditional correlationdetecting technology cannot guarantee the stability of gas-solid flow detection inharsh environments.In this paper, we construct a gas-solid flow parameter detection systemexperimental platform based on analysis of the parameters detection principle of thegas-solid flow, and the optical sensor is selected as detecting element. Analyzingcollected gas-solid flow signals in the time domain and frequency domain, the signalcharacteristic of time and frequency domain is found: enhancing the flow velocity, thepower between0and5kHz band increases and change frequency of signalaccelerates. According to the power spectrum analysis, the flow signal characteristicband which can characterize flow characteristics is determined. Then the signalcharacteristic is extracted and the parameter is obtained with the cross-correlationmethod. The experimental results show that the selected frequency band can becharacterized as a frequency characteristic of the flow characteristics. And because oflow correlation coefficient, low SNR and interference of correlated noise, thedetection performance of the traditional correlation detection method is not stableenough at low flow velocity.For the problem that when dealing with the signal of low correlation coefficient, low SNR and correlated noise interference, the traditional technology of gas-solidtwo-phase flow parameters detection gives no result, this paper presents a new methodof combining wavelet packet and singular value decomposition to process gas-solidtwo-phase flow signal. The method solves the problem of the wavelet packet analysisthreshold and threshold function selection. Aiming at difficulties of the singular valuedecomposition de-noising, an advanced algorithm of effective singular value selection,which based on variation of singular value curve angle, is presented in this paper. Byprocessing the simulation signal and the measured signal, the detection methods aresystematic comparative analyzed in this paper. The experimental results show that thismethod can still be able to achieve the parameter detection in the case of theconventional detection method unstable and increase the credibility of the test results.The method is more suitable for flow signal parameters detection of low signal noiseratio and correlated interference.