Parallel Finite Difference Simulation Of Cementing Quality Evaluation Method Using Obliquely Incident Ultrasonic Waves

Nowadays in domestic, the cement bond quality is evaluated by using some conventional methods, such as the CBL-VDL(Cement Bond Log and Variable Density Log). Although these methods are usually effective for evaluating cement bonding of the first interface(tube/cement interface), they are failed for the second interface(cement/formation interface). In this paper, numerical simulations are studied for the up-to-date method overseas of using obliquely incident ultrasonic waves. A series of analyses are present for the working principle of the method, for the criteria and results to evaluate the first and second interfaces, and for the design points of the apparatus.Firstly, the formulas for the velocity-stress staggered finite difference scheme are derived from the classical elastic dynamics equations. Based on this, a parallel finite-difference time-domain(FDTD) algorithm is implemented by Fortran 90 for simulating the borehole sonic waves. Then the borehole acoustic waves, especially those excited by obliquely incident ultrasonic waves in the borehole, are simulated by the algorithm. The main contents and results are as follows:The simple sequential algorithm is computationally expensive, thus it does not satisfy the research demand. In this study, an efficient parallel code is implemented by applying the MPI to improve the existing sequential FDTD algorithm in the cylindrical coordinate. The MPI technique and the workflow of the algorithm are introduced in Chapter 2. Furthermore, the correctness of the parallel code is illustrated by comparing the simulated acoustic waves of relatively simple models(axisymmetric uncased and cased boreholes) between the parallel and sequential algorithms.In Chapter 3, borehole ultrasonic waves of various models are calculated for the cement bonding evaluation of using obliquely incident ultrasonic waves, from which some of the significant problems are analyzed, such as the criteria for the cement bonding of the first and second interfaces. The wavelength of the ultrasonic waves with a central frequency of 250 kHz is much less than the borehole radius, thus the simulations can be simplified to the problem of 2D plan waves by neglecting the borehole effects. Thereby a 3D FDTD algorithm under the rectangular coordinate system is implemented in this study, which can degenerate for calculating the 2D problems. The computational efficiency by using this algorithm improves greatly compared with using the algorithm in cylindrical coordinate mentioned in Chapter 2. It is found that the amplitude and arrival time of the two wave groups, which are respectively referred to as the primary flexural wave and the secondary flexural wave, can be used as the criteria for the cement bonding of the two interfaces. Specifically, the amplitude of the primary flexural wave increase significantly when the first interface debonds and its attenuation coefficient decreases with the width of the debonding. In additional, the amplitude of the secondary flexural wave increase significantly when the second interface debonds and it arrives earlier with the width of the debonding. Besides that, this method is feasible for that of the most kinds of cements, except for the first interface bonding of the high impedance cements. At last, the wavefields for the declining apparatus are calculated. It is found that the method is failed when the angle of declining is more than ten degrees.In order to achieve an overall inspection, the cementing testing instrument is rotating at a high speed up to 8r/s when working in borehole. However, quality distribution and irregular instrument appearance will result in rotational imbalance and the tilting of instrument. This will cause the failure of measurement. Chapter 4 carried on the preliminary dynamic simulation on rotational balance, based on which the design points were given.