Three-dimensional geological modeling method

Since 1980s, researchers have put forward many 3D geological models to simulate geological bodies, and made great progress in this field. Through the research and analysis of a large number of domestic and foreign 3D geological modeling documents and professional software, 3D geological modeling methods can be roughly divided into three categories: discrete point source method, profile frame method and multi-source data coupling modeling method.

1.2.3. 1 discrete point source method

In geological prospecting, it is often necessary to obtain the shape of geological body according to a small amount of discrete sampling data (such as geological mapping or drilling data), which plays a guiding role in further guiding prospecting. Therefore, it is of certain significance to study how to realize the visualization of spatial scattered point data field.

Carlson (1987) proposed a three-dimensional conceptual model of underground space structure from the geological point of view, and proposed to establish a geological model with a simple complex model. Victor( 1993) and Pilout( 1994) applied the three-dimensional vector data model of Delaunay tetrahedron to study the geological modeling of discrete points. Lattuada( 1995) studied the application of 3D DT (three-dimensional Delaunay triangulation) in the geological field, and showed that tetrahedral mesh can be well used for three-dimensional modeling of geological bodies, and its advantages include: tetrahedral elements are easy to establish indexes; The model is easy to edit by hand; Topological structure can be derived from adjacent relations; Constrained triangulation is easy to realize surface constraint; Tetrahedron is very convenient for visualization and has high expression accuracy. It is easy to realize search and relational query. Courrioux et al. (200 1) realized the automatic reconstruction of geological target entities based on Voronoi diagram. Frank et al. (2007) used implicit functions to express 3D surfaces and reconstructed discrete points to simulate faults and salt domes. Yang Qin (200 1, 2005) used discrete point source information to construct strata and fault structural planes, and used them as constraints to constrain Delaunay subdivision to establish a three-dimensional geological model.

Borehole data also belongs to a kind of point source information. In essence, it effectively layers the original point and line data, and generates each level or entity by applying surface construction method according to the elevation of each level. Many scholars have studied 3D geological modeling around borehole data. Among them, Canadian scholar Houlding (1994,2000) used the information of borehole points for Delaunay triangulation as "primary tin", and other strata were realized by elevation mapping. Zhang Yu et al. (200 1) conducted in-depth research and development on its modeling method. In the ideal state of vertical drilling, a three-dimensional geological model is established by using the triangular prism data model, and the corresponding cutting algorithm is given. Lemon et al. (2003) established a three-dimensional stratum model by "stratum level method" and modified the geological model by user-defined cross section. Wu (2003) and et al. (2003) proposed a binary topological data structure modeling algorithm based on borehole data, and tried to construct underground three-dimensional geological model by using tetrahedral voxel model based on borehole data. Tetrahedral structure is more flexible in expressing complex structures, but using tetrahedron to represent spatial entities will produce a lot of redundancy, and the algorithm of generating tetrahedron is more complicated. Zhang Fang (2005) used Delaunay triangulation technology to construct a three-dimensional stratum model with borehole data, and introduced the idea of "interface segmentation" to adapt to the visual expression of massive data model, but it lacked the expression of geological body attribute information. Based on the triangular prism model, the quasi-triangular prism (ATP) (Qi, et al., 2002) and generalized triangular prism (Wu, 2004) methods are proposed to solve the problem of borehole deviation. At the same time, a similar triangular prism (STP)(Gong et al., 2004) was also proposed to solve the well deviation problem. For example, Kevin·Z et al. (2005) used STP to build a 3D geological model based on borehole data for virtual roaming of underground space. STP and GTP are essentially the same. Three-dimensional geological modeling based on borehole data, a seemingly simple data modeling method, has experienced the development of 10 years: from the initial TP data model to the ideal situation of vertical stratification of borehole and equal thickness of stratum, to the common situation that STP and GTP are suitable for non-vertical borehole and unequal thickness of stratum.

1.2.3.2 segmented frame method

Profile frame method is based on collecting and sorting out the original geological exploration data, establishing a classified database, manually generating a large number of two-dimensional geological profiles, and then generating a three-dimensional geological model of each horizon by surface structure method, or directly modeling geological bodies by voxel representation (Chae et al., 1999).

The main form of using geological profile expression to study regional three-dimensional geological phenomena is sequential geological profile method (Zhu Xiaodi et al., 200 1). The essence of sequential geological profile modeling technology is the computer realization of traditional geological mapping method, that is, the geological structure is described and geological information is recorded through plan or profile, as shown in figure 1.2. Its characteristic is that the three-dimensional problem is 2D, and several parallel or nearly parallel geological sections are used to express the spatial geological distribution characteristics of the study area, but the spatial expression is incomplete, and the stratigraphic or structural distribution between sections is left to engineers to "imagine". This modeling method is difficult to fully express the three-dimensional ore deposit and its internal structure.

Establishing true 3D model based on profile information has great development space and good adaptability to complex geological structure areas, and has become one of the main methods of geological modeling at present. However, 3D reconstruction based on contour has been improved and developed in the medical field, and rapidly expanded to other fields. In the medical field, a series of parallel human body slice images can be obtained by computer tomography (CAT) or magnetic resonance imaging (MRI), and a three-dimensional human body model can be generated by extracting the boundary of the object and based on the contour algorithm. Geological profile information, like medical profile information, reflects the structural distribution of the specific profile of the research object, and a three-dimensional geological model can be constructed with the help of medical three-dimensional human modeling technology. It is the application in archaeology that introduced the 3D modeling of medical profile into earth science earlier (Tipper, 1976,1977; Herbert et al., 1995), which is mainly used in the three-dimensional reconstruction of paleontology, but there are few documents about three-dimensional geological modeling.

Figure 1.2 Modeling example of sequence geological profile

The recognized masterpiece of 3D contour reconstruction is Keppel's article (Meyres et al.,1992; Herbert et al.,1995,2001; Xu et al., 2003; Qu Honggang et al., 2003). Meyres( 1992) based on Keppel's research, divided the section modeling method into four sub-problems: correspondence problem, tiling problem, branching problem and fitting problem: ① correspondence problem solves the contour matching problem between adjacent sections; (2) Network construction mainly solves the problem of triangle network construction between contours, and some criteria are considered, such as Keppel, 1975 and Fuchsetal. , 1977).③ The bifurcation problem is to solve the problem that the number of components of the same object on different sections is different; (4) Smoothing mainly solves the problem of interpolating the initially generated triangulation to get a smoother triangulation.

Qu Honggang et al. (2003) put forward the geological modeling method based on topological profile to realize the corresponding problems of complex geological three-dimensional modeling, and Deng Fei et al. (2007) discussed the general geological modeling of profile.

1.2.3.3 Multi-source data coupling modeling method

With the improvement of computer performance, people have the ability to process massive data, and the requirements for establishing geological models are getting higher and higher, hoping to establish high-precision and high-complexity geological models (Turner, 2003, 2006; Calcagno et al., 2006; Kaufman et al., 2008). Improving the accuracy of the model can be achieved by interpolation, but a better method is to refine the initial geological model by adding constraint information, which involves the problem of coupling multi-source data to establish a geological model.

As early as in 1993, when Holding put forward the concept of three-dimensional geoscience modeling, he emphasized that geological interpretation information has the function of modifying the model. It is pointed out that there are geological exploration data, manual drawing data and construction data in mining engineering, as well as uncertain data that need to be estimated by geostatistics (Houlding, 2000), and these different kinds of data need to be comprehensively considered in the final geological model.

McInerney et al. (2005a, b) think that 3D geological modeling can only be a digital geological sampling process, and more importantly, it is a manual interpretation process by geologists. Pointed out pointedly, don't expect some computer software to "model" automatically and successfully! It is realistic and necessary for an experienced geologist to input explanatory information for modeling. The software is just a convenient tool in the process of modeling (there is no expectation that some computer software will succeed and automatically "model"! In fact, explanatory input from skilled geologists is very important for modeling; The software is just a tool to simplify the modeling process. In fact, it is required that geological modeling should not only consider the deterministic data obtained from geological exploration, but also add the explanatory data of geological engineers to geological structures, which constitutes the basic idea of multi-source geological modeling.

Aiming at the particularity and complexity of geological body modeling, Mallet(2002) took point and line data as the main data sources, established a three-dimensional surface with triangle as the basic unit, and used Discrete Smooth Interdrift (DSI) to minimize the roughness of the surface. As the core technology of GOCAD, it has been supported by many geophysical companies and oil companies.

Compared with three-dimensional geological modeling with petroleum and mining engineering as the main application fields abroad, Zhong et al. (2006) started from the field of engineering geology of water conservancy and hydropower, studied multi-source geological data and established three-dimensional geological model of dam area. Wu et al. (2005) proposed a multi-source data integration geological modeling method with gradual subdivision. Considering the sparse geological data and low sampling rate, the initial geological model is continuously revised through gradual refinement.

The complexity of geological structure and the stage of understanding make multi-source geological modeling attract more and more research interests. The 32nd International Geological Congress (IGC) was held in Florence, Italy in 2004. On the topic of "the revival of geology" (Zanchi et al., 2007), many internationally renowned geoscience modeling experts * * * mentioned the problem of multi-source geological modeling. Among them, Zanchi et al. (2008) studied multi-source geological modeling of Italian Alps with the help of commercial software, and applied it to landslide stability analysis. Western developed countries mainly apply geological modeling to the fields of energy and environment, which is one of the few studies to open up new paths in the field of engineering construction. Coincidentally, Kaufmann et al. (2008) tried to use multi-source geological modeling to study the natural gas storage in abandoned coal mine roadway.

Generally speaking, 3D geological modeling technology is a development process from simple stratum simulation to complex geological structure simulation. From the initial establishment of a simple layered three-dimensional geological model based on a single data to the comprehensive utilization of multi-source data to establish a complex geological model, the spatial characteristics of geological structures can be reflected.