Zhuojin Pipe Fitting Co., LtdStudy Note on 15-Degree Finite Difference Serial Migration Offset and Compensation Methods
Literature Overview
This paper by Ling Yun, Zheng Yuying, and Guo Xiangyu from the Research Institute of Petroleum Geophysical Prospecting, published in Petroleum Geophysical Exploration (1989, Vol. 24, No. 6, pp. 645–659), addresses a fundamental problem in seismic migration: the over-migration artifact that occurs when applying 15° finite difference serial migration. While this topic falls outside the direct scope of steel pipe and welding engineering, the mathematical and computational methods discussed have indirect relevance to non-destructive testing (NDT) techniques such as ultrasonic tomography and phased array inspection, where wave equation migration principles are applied.
Core Technical Problem
The 15° approximation of the wave equation is widely used in seismic migration due to its computational efficiency. However, when applied serially (in multiple steps), the approximation error accumulates, leading to over-migration where reflectors are positioned at incorrect locations. This is analogous to systematic error accumulation in engineering measurement systems.
The paper identifies the following parameters that influence the differential error:
| Parameter | Symbol | Effect on Error |
|---|---|---|
| Time sampling interval | Δt | Larger Δt increases error |
| Spatial sampling interval | Δx | Larger Δx increases error |
| Depth step | Δτ | Larger Δτ increases error |
| Layer velocity | V | Higher velocity reduces error |
| Frequency | f | Higher frequency increases error |
| Interface dip angle | φ | Steeper dip increases error |
| Second-order approximation coefficient | B | Depends on discretization scheme |
Compensation Method
The authors propose introducing a compensation coefficient R into the original finite difference wave equation. The coefficient R is a function of all the parameters listed above:
R = f(Δt, Δx, Δτ, V, f, φ, B)
By incorporating R, the method corrects for the cumulative differential error while preserving the computational advantages of the 15° approximation. The study demonstrates that with proper R selection, interfaces with dips up to 60° can be correctly positioned.
Relevance to Engineering Practice
While this paper addresses seismic exploration, the underlying principle of error compensation in wave propagation modeling has direct parallels in ultrasonic NDT. In phased array ultrasonic testing (PAUT) of welds and pipe joints, wave equation-based migration techniques are used to reconstruct defect images from raw scan data. The same type of systematic error accumulation can occur when processing ultrasonic data in layered or anisotropic materials, such as welded steel pipes with grain texture.
Engineers working with advanced NDT techniques should be aware that:
- Sampling parameters (spatial and temporal) must be carefully selected to minimize discretization error
- Steeply oriented defects (high dip angle relative to the scan plane) are more susceptible to positioning errors
- Compensation algorithms can improve defect localization accuracy without sacrificing computational efficiency
Key Questions and Reflections
The paper does not address the effect of anisotropy on the compensation coefficient, which is a significant limitation for applications in textured welded materials. Additionally, the method assumes a constant velocity model, whereas real materials exhibit velocity gradients, particularly in heat-affected zones of welds.
This study is a valuable reminder that mathematical approximation errors must be systematically quantified and compensated in any wave-based engineering method. For NDT practitioners, understanding the relationship between sampling parameters and positioning accuracy is essential for reliable defect characterization.