Studies Of Oil Holdup Measurement For Oil-water Two Phase Flow Based On Thermal Method

Oil-water two phase flows are a common occurrence in petroleum industry, chemical engineering, power and process industries, in which the oil holdup is one of the important flow parameters. The determination of the oil holdup has great significance to the measurement and control of oil flux and water flux respectively. Because the flow structures of oil-water two phase flow (OWTPF) are very complex, therefore, accurate measurement of the oil holdup is very difficult. The oil holdup measurement methods are researched on according to the fluctuant temperature signals of OWTPF measured by thermocouple and temperature field transformation measured by platinum resistance, which is based on the thermal method and modern signal processing techniques. The mainly research works of this thesis are as following:1. A new signal processing method based on Multiwavelet has been proposed for temperature signal de-noising of OWTPF for the first time. Compared with traditional wavelet de-noising, the Multiwavelet Neighboring Coefficient Method can deal with the correlativity among the transformation coefficients and the actual temperature signal is effectively retained, the measurement range of oil holdup is improved. A new model based on least square support vector machines (LSSVM) and multiwavelet transform is put forward on, which is capable of forecasting oil holdup of OWTPF. The fluctuant temperature amplitude signals noised and standard difference are employed as the characteristics parameters of the oil holdup, the LSSVM model of oil holdup is built. The experimental results demonstrate this model has good measurement accuracy.2. A new measurement method of oil holdup for oil-water two phase flow is put forward on the base of Hilbert-Huang translation (HHT) and LSSVM. The temperature signals of OWTPF are decomposed by the HHT. The instantaneous fluctuated amplitude that relates to oil holdup is extracted form the last residue component of HHT. Meanwhile the high frequency components are analyzed by the FFT(Fast Fourier Transformation), it is found that the FFT maximum energy value is relates to oil holdup. Therefore, the maximum fluctuant amplitude and FFT energy are employed as the characteristics parameters of LSSVM model, the LSSVM model of oil holdup is obtained based on HHT. The experiment results show this model can be used for oil holdup measurement of OWTPF with reasonable accuracy. 3. In order to improve the predictive accuracy and generalization ability of the LSSVM model, a Genetic Arithmetic (GA) and Particle Swarm Optimization (PSO) has been adopted to determine the optimal parameters of LSSVM model automatically, which utilize their global searching. The optimal model of oil holdup is obtained.4. Oil holdups model of oil-water two phase flow is developed utilizing platinum resistance to measure the temperature difference between upstream and downstream based on the fluid thermal balance equation. The measurement errors of oil holdup are analyzed and measurement method is improved on. The calibration coefficient of temperature difference and flux is proposed to calibrate temperature difference and oil water total flux measured. The experimental results show that the average measurement error is 4.56% in the range of 0 to 60% oil holdup. Subsequently, the LSSVM model of oil holdup is built and is optimized using GA and PSO, the measurement average errors are less than 1%, which reveals that the performance of LSSVM model for oil holdup outperforms the traditional model.