Productivity Prediction Of Hydrate Reservoirs In Shenhu Area

Natural gas hydrate resources are huge and potential clean energy in the 21st century.With the increasing production test of gas hydrate reservoirs worldwide,it is of great significance to predict the gas production capacity of gas hydrate reservoirs.Reasonable productivity prediction will provide very important guidance for the production test and development of gas hydrate reservoirs.In this paper,analytic solution,similarity theory and numerical simulation methods are used to predict the productivity of hydrate reservoirs.Secondly,taking the productivity of hydrate reservoirs as the evaluation criterion,the quality level of hydrate reservoirs in Shenhu sea area is predicted based on BP neural network method.Based on the characteristic that hydrate decomposition provides additional gas source,the material balance equation of hydrate reservoir is deduced.Combined with energy conservation equation and inflow dynamic equation,the analytical solution formula of productivity is deduced.The correctness of the analytical solution is verified by comparing the results of analytical solution and numerical solution.The maximum decomposition front expression under different formation conditions is deduced under the steady state condition of hydrate reservoir.It is found that the closer the formation pressure is to the phase equilibrium pressure,the lower the bottom hole pressure,the larger decomposition zone permeability,the smaller caprock permeability,the larger the hydrate saturation the thicker the hydrate layer,the smaller the caprock thickness and the larger the maximum decomposition front.Under the same production pressure,the radius of the maximum decomposition front in horizontal well production is larger than that in vertical well production,and the smaller the bottom hole pressure is,the larger the difference of the maximum decomposition front between horizontal well production and vertical well production.Based on the mathematical model of HRS software,71 similarity criteria are deduced by using similarity theory.Through the analysis of similarity scaling,it is found that the ratio of injection and production speed is 7/3 power of geometric scaling and the ratio of time scaling is 2/3 power of geometric scaling.According to the similar shrinkage ratio,the experimental results are enlarged,and the actual hydrate reservoir productivity prediction is realized.Through numerical simulation,the pressure reduction performance of vertical and horizontal wells in Shenhu hydrate reservoir is compared.The results show that the pressure reduction effect of horizontal wells is better.Through the optimization of horizontal wells,it is found that horizontal wells are more conducive to mining in the upper hydrate layer.The greater the formation penetration ratio,the better the mining effect.But when the penetration ratio is greater than 0.8,the mining effect is basically the same.By comparing the results of magnification and numerical simulation,it is found that the results of numerical simulation method are lower than those of similar magnification method,but the gas production is in the same order of magnitude.The productivity of hydrate reservoirs in Shenhu area is predicted by numerical simulation and analytical method.The result shows that the cumulative gas production in 10years is 1.25×10⁸ m³ by numerical simulation and 1.25×10⁸ m³ by analytical method.By calculating the sensitivity of hydrate reservoir parameters,the results show that the parameters with sensitivity magnitude above 10^-2 are formation temperature,rock thermal capacity,formation porosity,hydrate density,hydrate solid concentration,formation thickness,formation pressure,formation permeability,rock thermal conductivity,water thermal conductivity,gas thermal conductivity,formation compressibility and initial water saturation.With these parameters as the input of the neural network model and productivity as the evaluation criterion,the BP neural network model is constructed.The learning accuracy of the model after training is 93.3%.The model is used to evaluate the quality of hydrate reservoirs.The results show that the hydrate reservoirs in the South China Sea Trough of Japan are grade1 hydrate reservoirs,and in the Shenhu Sea area are grade 3 hydrate reservoirs.