阅读说明：本技术 一种基于fpga的地震正演计算方法 (Earthquake forward modeling calculation method based on FPGA ) 是由 田志明 高雨豪 于 2021-09-26 设计创作，主要内容包括：本发明涉及到油气勘探技术领域,尤其涉及到一种基于FPGA的地震正演计算方法,本发明将FPGA封装成为一个通用的波场延拓计算的算子库,通过在CPU侧进行数据计算的过程进行设置,将整个过程按照不同的需求进行波场延拓。在进行逆时偏移的过程中,将两个过程用同一个FPGA计算的bit计算,将计算的波场值记录下来,在另一过程需要计算反向过程时,将参数设置为反向延拓的状态计算出反向的波场值,并将波场值记录。本发明利用同一个bit进行地震波场正演正演计算,在整个测试过程中也充分发挥了FPGA的并行计算的优势,同时大大降低了FPGA开发的工作量。(The invention relates to the technical field of oil-gas exploration, in particular to an FPGA-based seismic forward modeling calculation method. In the process of reverse time migration, the two processes are calculated by using bits calculated by the same FPGA, the calculated wave field value is recorded, and when the other process needs to calculate the reverse process, the parameters are set to be in a reverse continuation state to calculate the reverse wave field value, and the wave field value is recorded. The invention utilizes the same bit to carry out forward calculation of the seismic wave field, fully exerts the advantages of parallel calculation of the FPGA in the whole test process and greatly reduces the workload of FPGA development.)

1. An FPGA-based seismic forward modeling calculation method is characterized by comprising the following steps:

step S1: the CPU loads a data document and transmits the data document into the FPGA after the data document is loaded;

step S2: performing wave field continuation calculation in the FPGA by adopting a finite difference method according to the wave field continuation process;

step S3: judging whether the wave field value is a recording moment or not according to a preset snapshot recording interval, if the wave field value calculated by the FPGA needs to be recorded, transmitting the wave field value to a CPU (central processing unit) through a PCIE (peripheral component interface express) interface, and generating the wave field value in a disk at the CPU side and recording a wave field snapshot; if not, continuing the next step;

step S4: determining whether wave field propagation is forward or backward, updating iteration time, deducing time backward if the wave field propagation is forward wave field expansion, and loading time backward if the wave field propagation is backward wave field expansion;

step S5: and repeating the steps S2-S4 until the whole wave field continuation is completed.

2. The FPGA-based seismic forward modeling calculation method of claim 1, wherein the data file comprises velocity field data, observation system data, and wave field data.

3. The FPGA-based seismic forward modeling calculation method of claim 2, wherein in step S4, if the forward wavefield is expanded, the wavefield value of the corresponding wave detection point at that time is recorded as a wavefield record according to the observation system data, and the time is deduced backwards.

4. The FPGA-based seismic forward modeling calculation method of claim 2, wherein in step S4, the wave field propagation is backward wave field propagation, the wave field value corresponding to the wave detection point at that time is loaded into the wave field according to the observation system data, and the time is loaded backward.

Technical Field

The invention relates to the technical field of oil and gas exploration, in particular to an FPGA-based seismic forward modeling calculation method.

Background

Along with the deepening of the oil-gas exploration degree, the difficulty of the oil exploration is increased, the target of the oil-gas exploration is changed from a large-scale structured oil-gas reservoir to a small-scale complex hidden oil reservoir, so that the improvement of the lithology recognition and fluid prediction precision is important in actual production, the acquisition precision is high, and the significance of indicating the lithology and the geophysical parameters of the fluid is great.

At present, in the face of complex exploration targets, how to improve inversion accuracy and obtain more accurate reservoir parameters is particularly important. Full Waveform Inversion (FWI) techniques utilize kinematic and kinetic information of a prestack seismic wave field to quantitatively extract parameters of an underground medium and reconstruct an underground velocity structure, and can provide reliable velocity data for prestack depth migration, lithology judgment and reservoir identification, however, the huge forward calculation requirements and the high nonlinearity of a target functional of a Full waveform inversion method prevent the Full waveform inversion method from being widely applied to seismic exploration, and particularly, the Full waveform inversion based on a multi-component wave equation has insurmountable calculation amount, memory space and calculation time.

The ultra-large calculated amount and the storage amount are also main factors for restricting one-step development of the full waveform inversion, different scholars perform extensive research for relieving the calculation and storage pressure, because wave equations need to be solved in forward propagation of a seismic source wave field and backward propagation of wave field residual errors, the calculation time of the full waveform inversion is closely related to the forward modeling algorithm, and therefore the research of the fast and efficient forward algorithm has great significance for improving the calculation efficiency of the full waveform inversion.

Like reverse time migration, time domain full waveform inversion suffers from the problem of large memory, and this method extends the source wavefield in the forward direction along time, and requires backward access to the data along time when calculating the parameter gradients. Therefore, wavefields at all times in all grids need to be stored, which is clearly unacceptable for larger geological models. A simple solution is to output the wavefield to disk and read from disk when needed, but the disk input and output itself is quite time consuming, especially for high performance computers, disk reads and writes are much more expensive than floating point type operations.

In recent years, with the improvement of the level of computer hardware, the geophysical technology is not continuously developed, and a full waveform inversion method is also rapidly developed, and the method gradually draws more and more attention from geophysical workers, and becomes one of the research hotspots in the field of geophysical exploration. Full waveform inversion seismic imaging is a seismic imaging technique based on data fitting objective function optimization, which can fully utilize all information (amplitude, phase, etc.) of the wavefield, and thus has the potential to reconstruct high resolution seismic images. In addition, wave equation forward simulation is a basic unit of full waveform inversion, a forward simulation algorithm has great influence on the accuracy and efficiency of waveform inversion, and the research on a wave equation forward algorithm with high simulation accuracy, high speed and high efficiency is also very important.

For a long time, the improvement of computing performance is mainly benefited by the improvement of the working frequency of a microprocessor, and Moore's law dominates the pace of the development of information technology. However, the problem of system heat generation due to the increase of the main frequency is becoming more and more prominent, and the system power consumption is increasing, so in recent years, new ideas are adopted for the development of CPU technology, and the pace of increasing the main frequency is slowed down or even stopped, and the improvement of the parallel processing capability of the processor is becoming a main technical idea. In order to ensure the continuous improvement of the performance of a microprocessor chip and more importantly to reduce the power consumption and complexity of the chip, the current mainstream commercial CPU design completely adopts a multi-thread multi-core architecture, dual cores and four cores become mainstream products of the CPU, six-core and eight-core CPU products are also produced in large scale in batch, and the continuous increase of processor cores (called as a many-core processor) is expected to become a main characteristic of CPU technical development for a period of time in the future. How to develop the multi-core parallel computing program quickly and effectively is important for giving full play to the performance of a multi-core processor system. Besides the technological route of traditional CPU from single core to multi-core (many cores), several new technological routes are currently active, which may represent an important direction for future high-performance computing technology development. One important direction is reconfigurable computing technology based on FPGA (field programmable gate array), and the heterogeneous parallel computing system architecture is adopted by the petascal computer system representing the highest computing level at present, both internationally and domestically. How to construct a multi-level parallel computing software development framework and a programming tool on the heterogeneous parallel computing system and promote the development and transplantation of large-scale parallel computing application software is the key for realizing the large-scale popularization and application of the heterogeneous parallel computing system.

Disclosure of Invention

In view of the above technical problems, the present invention provides an FPGA-based seismic forward calculation method; the method greatly increases the application and popularization of the FPGA in seismic exploration by utilizing the parallel computing advantage of the FPGA, and simultaneously provides a feasible solution for solving the problem of seismic exploration.

The technical scheme adopted by the invention for solving the technical problems is as follows:

an FPGA-based seismic forward modeling calculation method is characterized by comprising the following steps:

step S1: the CPU loads a data document and transmits the data document into the FPGA after the data document is loaded;

step S2: performing wave field continuation calculation in the FPGA by adopting a finite difference method according to the wave field continuation process;

step S3: judging whether the wave field value is recorded according to the preset snapshot record interval

If the wave field value calculated by the FPGA needs to be recorded and transmitted to a CPU (central processing unit) through a PCIE (peripheral component interface express) interface, generating the wave field value in a disk at the CPU side and recording a wave field snapshot; if not, continuing the next step;

step S4: determining whether wave field propagation is forward or backward, updating iteration time, deducing time backward if the wave field propagation is forward wave field expansion, and loading time backward if the wave field propagation is backward wave field expansion;

step S5: and repeating the steps S2-S4 until the whole wave field continuation is completed.

The seismic forward modeling calculation method based on the FPGA is characterized in that the data document comprises velocity field data, observation system data and wave field data.

The seismic forward modeling calculation method based on the FPGA is characterized in that in step S4, if the forward wavefield is expanded, a wavefield value corresponding to the wave detection point at the time is recorded as a wavefield record according to observation system data, and time is deduced backwards.

The seismic forward modeling calculation method based on the FPGA is characterized in that in the step S4, wave field propagation is expanded into a reverse wave field, a wave field value of a corresponding wave detection point at the moment is loaded into the wave field according to observation system data, and time is loaded backwards.

The technical scheme has the following advantages or beneficial effects:

according to the seismic forward modeling calculation method based on the FPGA, the common seismic forward modeling is packaged into an operator library for FPGA calculation according to the calculation process of reverse time migration imaging and the characteristics of full waveform inversion. The method analyzes the requirement of large calculation amount in the seismic exploration calculation process by utilizing the parallel calculation advantage of the FPGA, packages the forward evolution of the seismic wave field, and calculates the wave field according to different observation systems in different calculation processes. The invention utilizes the same bit to carry out forward calculation of the seismic wave field, fully exerts the advantages of parallel calculation of the FPGA in the whole test process and greatly reduces the workload of FPGA development. The application and popularization of the FPGA in seismic exploration are greatly increased through the invention, and a feasible solution is provided for solving the problem of seismic exploration.

Drawings

The invention and its features, aspects and advantages will become more apparent from reading the following detailed description of non-limiting embodiments with reference to the accompanying drawings. Like reference symbols in the various drawings indicate like elements. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.

FIG. 1 is a flow chart of an FPGA-based seismic forward modeling calculation method.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The invention packages the general forward process into a general module, extracts the forward process of the wave field continuation full waveform inversion of reverse time migration into an independent module through the parameter setting of the CPU side, and adopts the finite difference calculation equation to carry out the wave field continuation according to different calculation processes. And performing calculation by using the FPGA to perform forward modeling, loading a seismic source at a shot point according to the shot-geophone relation of the observation system, and recording the seismic wave field record at a demodulator probe. When the process of reverse continuation is carried out, the recorded data is taken as a seismic source and loaded into a wave field according to the appointed time point according to the position of a wave detection point; in the whole process, wave field snapshots are recorded according to different moments. The specific implementation process is shown in a forward simulation flow chart in FIG. 1:

step S1: the CPU loads a data document and transmits the data document into the FPGA after the data document is loaded;

step S2: performing wave field continuation calculation in the FPGA by adopting a finite difference method according to the wave field continuation process;

step S3: judging whether the wave field value is a recording moment or not according to a preset snapshot recording interval (Nsnap time recording step length), if so, transmitting the wave field value calculated by the FPGA to a CPU (central processing unit) through a PCIE (peripheral component interface express) interface, and generating the wave field value in a disk and recording a wave field snapshot on the CPU side; if not, continuing the next step;

step S4: determining whether wave field propagation is forward or backward, updating iteration time, deducing time backward if the wave field propagation is forward wave field expansion, and loading time backward if the wave field propagation is backward wave field expansion;

step S5: and repeating the steps S2-S4 until the whole wave field continuation is completed.

In an embodiment of the invention, the data file contains velocity field data, observation system data, and wave field data. In step S4, if the forward wave field is expanded, the wave field value corresponding to the wave detection point at that time is recorded as a wave field record according to the observation system data, and the time is deduced backwards. And the wave field propagation is a reverse wave field expansion, the wave field value corresponding to the wave detection point at the moment is loaded into the wave field according to the data of the observation system, and the time is loaded backwards.

In the whole calculation process, the FPGA is packaged into a general operator library for wave field continuation calculation, and the wave field continuation is carried out in the whole process according to different requirements by setting the calculation process on the CPU side. In the process of reverse time migration, the two processes are calculated by using bits calculated by the same FPGA, the calculated wave field value is recorded, and when the other process needs to calculate the reverse process, the parameters are set to be in a reverse continuation state to calculate the reverse wave field value, and the wave field value is recorded. And finally, completing subsequent imaging calculation according to imaging conditions, and calculating by using the same FPGA-encapsulated bit in the process of calculating the total wave inversion. Meanwhile, different earthquake forward process calculations can be carried out according to different calculation requirements.

The method analyzes the requirement of large calculation amount in the seismic exploration calculation process by utilizing the parallel calculation advantage of the FPGA, packages the forward evolution of the seismic wave field, and calculates the wave field according to different observation systems in different calculation processes. The invention utilizes the same bit to carry out forward calculation of the seismic wave field, fully exerts the advantages of parallel calculation of the FPGA in the whole test process and greatly reduces the workload of FPGA development. The application and popularization of the FPGA in seismic exploration are greatly increased through the invention, and a feasible solution is provided for solving the problem of seismic exploration.

Those skilled in the art will appreciate that those skilled in the art can implement the modifications in combination with the prior art and the above embodiments, and the details are not described herein. Such variations do not affect the essence of the present invention and are not described herein.

The above description is of the preferred embodiment of the invention. It is to be understood that the invention is not limited to the particular embodiments described above, in that devices and structures not described in detail are understood to be implemented in a manner common in the art; those skilled in the art can make many possible variations and modifications to the disclosed embodiments, or modify equivalent embodiments, without affecting the spirit of the invention, using the methods and techniques disclosed above, without departing from the scope of the invention. Therefore, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention are still within the scope of the protection of the technical solution of the present invention, unless the contents of the technical solution of the present invention are departed.