Research On The Novel Transit Time Measurement And Data Integration Of Multi-path Ultrasonic Gas Flow Meter

With the development of gas market in China,accurate and efficient gas flowmwter is urgently needed for gas trade.The ultrasonic gas flowmeter,with the advantages of wide range ratio,low pressure loss,no baffle,high accuracy of large diameter flow measurement,and convenience for pipeline cleaning,has a broad application prospect in the field of gas measurement and trade.Besides,the multi-path ultrasonic flowmeter can detect the distribution of gas flow field more completely compared with the single path one,so multi-path ultrasonic gas flowmeter is more accurate.However,due to the inadequate research in China,high-precision multi-path ultrasonic gas flowmeter still relies on import.Therefore,it is of great importance to develop domestic high-precision multi-path ultrasonic gas flowmeter.The transit-time-difference-based multi-path ultrasonic flowmeter is widely used owing to its simple principle,whose key technique is the measurement of the ultrasound transit time.The two typically adopted methods applied for transit time acquisition are threshold method and cross correlation algorithm.The first method is simple but prone to trigger wrongly when shape distortion of waveform occurs.The latter method is based on the similarity of overall received waveform,so it has stronger anti-interference ability than the former.However,limited by the sampling frequency of the waveform,the resolution of cross correlation algorithm is lower than threshold method.In addition,the accuracy of multi-path ultrasonic flowmeter is also affected by the data integration algorithm.Traditional data integration algorithms with fixed weights are sensitive to the velocity profiles of flow,which will result in errors for the calculation of section average velocity.Therefore,improved transit time measurement method and data integration algorithm are studied in this thesis.The main work and innovations of the thesis are listed as follows:1.An improved cross correlation algorithm based on reference waveform update is proposed.The correlation coefficient between the measured waveform and the reference waveform is calculated.The weighted summation of the measured waveform with high correlation coefficient and reference waveform is used as updated reference waveform.Meanwhile the weight coefficient changes adaptively with the correlation coefficient.The algorithm makes the reference waveform adaptive to the vibration mode change of the ultrasonic waveform,which improves the stability of transit time measurement at different temperatures.2.In order to improve the resolution of the transit time,zero crossing detection is combined with cross correlation algorithm.The basis of this algorithm is that the onset time difference of two waveforms can be reflected by the zero crossing position,although the time difference is less than one sampling period.The starting cycle of the received waveform is determined by the cross correlation algorithm,and then the zero crossing point behind the starting cycle with the minimum phase difference is selected as the feature point for determining the transit time.Sampling points near the feature point is interpolated by cubic spline,and the abscissa of the feature point is calculated by the linear fitting of two interpolation points whose amplitude is closest to zero.The transit time is calculated based on the abscissa of the feature point,so the difference of the transit time less than one sampling period can be detected.Therefore,the accuracy of transit time in small velocity is improved,although the sampling frequency is limited.3.The error of traditional numerical integration algorithm applied is studied,and an improved numerical integration method with variable weight function based on velocities of acoustic paths is proposed to calculate the pipeline section average velocity.In this thesis,the average velocity of pipeline section is estimated based on the approximate linear relationship between the velocity of each path and the average velocity of the pipeline section.Combined with the ideal flow velocity distribution model,the current velocity distribution function is predicted,which is used to update the weight function of numerical integration.The updated weight function is more consistent with the actual flow velocity distribution,which can reduce the integration error.4.The gas flow experiments are conducted to verify the effectiveness of the proposed methods using the relative error,repeatability and generalization as performance indicators.The experiment results demonstrate that the proposed transit time calculation method is capable of improving the resolution of transit time,which improve the measurement accuracy of small velocity.Additionally,the comparison shows that the relative error of the improved numerical integration method proposed is less than traditional Gauss-Legendre numerical integration algorithm.Moreover,the repeatability and generalization performance of improved numerical integration method is better than that of the least squre regression method.After the real flow correction,the relative error of the flowmeter is basically less than ±1%and the repeatability is basically less than 0.4%in the range.