Experimental Study On Hydrate Saturation Model Based On Electrical-Acoustic Characteristic Parameters

Natural gas hydrate has become a great potential strategic energy and a global energy research hotspot,because of its large reserves,clean and pollution-free characteristics.Hydrate saturation is one of the important parameters for quantitative evaluation of natural gas hydrate reservoir.Accurate calculation of hydrate saturation has guiding significance for hydrate exploration and development.Clay content and hydrate saturation in natural gas hydrate sediments can significantly affect the electrical and acoustic properties of sediments.Taking impedance modulus and acoustic velocity as input variables and hydrate saturation as output variables,a comprehensive calculation model of hydrate saturation is established,which is of great significance for quantitatively evaluating gas hydrate reservoir by using resistivity logging and acoustic logging data.The existing hydrate sediment test equipment was upgraded and the hydrate formation and decomposition experiments were carried out in sediment system with different clay content.The electrical,acoustic,temperature and pressure data in the process of hydrate formation and decomposition were obtained.The measured data were preprocessed,and the impedance,phase angle,sound velocity and other characteristic parameters were calculated.The influence of saturation and clay content on Electrical and acoustic characteristic parameters is studied.The results show that:(1)with the increase of hydrate saturation,the impedance modulus and sound velocity show a non-linear growth trend;(2)when the clay content in the sediment is less than20%,the additional conductivity of clay minerals enhances the overall conductivity of the sediment;when the clay content is more than 20%,the water swelling of clay minerals produces a film effect in the pores,which reduces the migration rate of conductive ions in the pore water,thus weakening the overall conductivity of the sediment.In view of the small amount of data obtained from THF hydrate experiment,support vector regression(SVR)was used to establish four hydrate saturation calculation models based on broadband(20Hz ～ 100 k Hz)complex resistivity modulus,phase angle,real part and imaginary part data.The weighted average fusion strategy is used to fuse the single saturation model,and the SVR hydrate saturation calculation model with multiple complex resistivity parameters as input is established.The influence of clay content on the prediction effect of SVR hydrate saturation model is analyzed.Based on the broadband complex resistivity parameters of sediment,the equivalent circuit model and Archie formula model are established.The SVR model is compared with the equivalent circuit model and Archie formula model by using mean relative error(MRE),mean absolute error(MAE)and root mean square error(RMSE)as indexes to evaluate the prediction effect of the three models.The results show that the model proposed in this paper has rich input information,reduces the error caused by empirical parameters.Its prediction effect is better than the other two electrical models,and its error of MRE,MAE and RMSE is less than 4.0%,0.012% and 5.0% respectively.Based on a large number of electrical / acoustic characteristic parameters obtained from methane hydrate experiments,the hydrate saturation calculation model is established based on the Integrated BP neural network(BPNN-Ada)and the saturation calculation formula of conventional resistivity method and sound velocity method.MRE,MAE and RMSE are used as indexes to evaluate the prediction effect of the three models.The results show that: the hydrate saturation calculation model based on the electroacoustic characteristic parameters has more advantages than the conventional empirical formula and theoretical model in dealing with the complex mapping relationship between the electroacoustic parameters and the hydrate saturation and the Integrated BP network saturation model has higher prediction accuracy,its three error indexes are MRE = 0.48%,MAE = 0.0005 and RMSE = 0.76% respectively;the input information of the model is more abundant,and the electrical / acoustic characteristic parameters play a complementary role,which enhances the reliability and applicability of the model;by using a suitable fusion strategy to integrate the electroacoustic measurement data of multiple sensors in different measurement directions,the adverse effect of uneven hydrate spatial distribution on the accuracy of saturation calculation can be overcome.