A New Method for Rapid Identification of Tight Oil Reservoir Fluid Property Logging in the Foreland Basin on the Southwestern Margin of Ordos Basin

Abstract

As known, rock drillability and interval acoustic transit time both are important data for studying the physical properties of assise. Previous studies have proposed the relationship between logging and mud-logging data, and it has been verified that the inverse correlation between rock drillability and interval transit time exist. Generally speaking, the rock skeleton with dense formation has poor drillability and fast speed of acoustic propagation, while the pore of rock is opposite. Therefore, the stronger the drillability of rock is, the greater the time difference of sound wave is (Lou 1997). Logging data filtering usually face multiple problems effected by well conditions, such as random error, Non-geological factors and so on (Cui 1995). The previous researches have attempted to overcome the numerical random distortion problem by using Kalman filtering (Tang 2004). In this paper, an acoustic logging filtering method based on corrected ROP(rate of penetration) is proposed to improve the above problems to some extent. The filtering scheme considers that the mud-logging represents physical dynamic characteristics of drill hole and acoustic logging gives an index to the static physical properties of adjacent rock. Based on simplified Kalman filter, our team establishes an integrated state parameter to approximate the real response of the system. The predict parameter is established according to corrected ROP as the predicted value, while acoustic time difference is set as the measured value, so as to achieve the filtering equation for realizing borehole compensation and data smoothing of acoustic logging. A list of the process flow is roughly shown as follows: firstly, the original ROP is corrected by drilling engineering or mud-logging parameters to obtain the correct ROP for eliminating the influence of non-geological factors. Then, the status-prediction equation of drilling footage and rock drillability are established to adjust parameter for the update equation based on both acoustic and ROP data. At last, the status-prediction and the updating equations are solved in parallel to realize filtering. The experimental results show that this method is effective enough to be applied and operated in mud-logging field drilling easily. In addition, previous studies have shown that, usually, the quantitative relation between the volume of the log gas can characterize the fluid property of reservoir to a certain extent. The proportion of light component in oil-produced reservoir in the study area is much higher than the average level of onshore reservoirs in China, which makes it more difficult to judge the oil-gas content of the reservoir by traditional methods. The author’s team introduced time-depth transformation and relative rate of change modeling to reconstruct logging gas component data, forming a rapid identification method for nCs reservoir fluid. The method calculates the relation between the relative contents of ethane with propane and isomeric butane with isomeric pentane, and forms the oil index curve and the water risk index curve which represent the oil content of the reservoir and the possibility of water content.