I think it may be a good idea to explain why Acceleware is so excited about our seismic imaging roadmap. It is a truism in the petroleum industry that “all the easy oil has been found.” New meaningful reserves will only be discovered using innovative technology and techniques. Acceleware is assisting the world’s largest oil & gas players find new oil with its seismic imaging products. So how exactly does this work?
Seismic data acquisition and processing play a key role in oil & gas exploration. The aim is to use seismic data to provide a greater understanding of sub-surface geological structures and, hence, to determine their suitability for trapping valuable hydrocarbon reservoirs. Essentially, it’s used to decide where to drill wells. The cost (and risk) of drilling dry holes particularly off-shore is becoming astronomical. Add that to the fact that we are now searching for “hard” oil leads to the requirement for advanced seismic processing.
For the entire history of seismic processing, the ability to model the Earth physics has been limited by the computational power available at the time. Until relatively recently, seismic processing techniques were rather simplistic, necessarily relying on gross simplifying approximations purely because anything more complex was beyond the scope of existing computational resources. In the days of easy oil, conventional seismic processing was sufficient. However, the “hard” oil is now in complex geological formations (sub-salt, shale, sub-basalt, etc). Places where conventional seismic techniques have been challenged to successfully image. But now, advances in computer architecture over the last few years - particularly GPU computing, combined with improved algorithmic techniques have made it possible to process seismic data in much more accurate and realistic ways, honouring the physics of seismic wave transmission and reflection to unprecedented degrees. This is how the “hard” oil is going to be found.
How will Acceleware help? First we need to understand how oil & gas companies use seismic data. At the risk of oversimplifying the process, here is how it works. Someone determines that a particular geological formation is a candidate for exploration, but in order to determine where to drill a well an accurate 3D geological model is required. A decision is made to “shoot some seismic” – that is design a seismic acquisition program (a survey) that will provide clues to the underlying geology. Once the seismic survey is designed, the data is acquired in the field and then processed. At first the data is processed in a fairly conventional sense to produce a seismic image. Processing requires that the geophysicist already has a model of the earth. Usually the initial models are crude approximations, so processing is performed iteratively with the model being updated each time. At some point a decision is made that more advanced processing is required to refine the model in an area of interest. One such advanced method is Reverse Time Migration (RTM). Once an accurate image is produced a geophysicist must interpret the image and build a full 3D geological model which can be used to pick an exploration target and drill a well. The results of the well are used to confirm the 3D model at that point. The company may then decide to produce the well, acquire more data, drill more wells, or abandon the prospect. Acceleware assists oil & gas companies by developing advanced seismic imaging methods and techniques that reduce the time invested in this inherently iterative process.
At the core of any new advanced seismic processing method is accurate modeling of seismic wave propagation through the Earth. Thus, whereas many of the older methods rely on viewing seismic energy travelling along discrete rays (very often in straight lines), the newer methods use detailed physics to accurately model propagation of a full 3D seismic wavefield, accounting for propagation in all directions, as well as proper representation of reflection and refraction of the wavefield through the varying rock properties of the sub-surface.
Such detailed physics-based wavefield propagation modelling is at the center of Acceleware’s industry-leading Reverse Time Migration product (AxRTM™) and its derivatives. At its core, the AxRTM library is a high performance wave propagation engine. That engine can be used to generate highly accurate simulations of seismic field experiments (forward modelling), perform RTM imaging, and will soon provide Full Waveform Inversion (FWI) capabilities. RTM is very much the current state-of-the-art in seismic imaging, overcoming many of the limitations of simpler imaging techniques, providing the clearest possible image of even the most complex geological environment. The standalone forward modelling mode of AxRTM is very useful in seismic survey design, understanding the interaction of the wavefield with complex geology, and in interpretation of seismic data.
The current roadmap for Acceleware seismic products will further enhance the company’s position at the leading edge of new seismic imaging technologies by building on AxRTM in two different directions. The first direction will be to include modeling of full elastic wavefields. Almost all seismic imaging done today assumes that seismic waves are purely compressional, i.e. they behave like acoustic, or sound waves. However, full seismic wavefields also include so-called shear waves, which may carry additional valuable information, for example the fluid content of the geology. Analysis of these shear waves requires full elastic modeling, which is set to become a huge growth area within the seismic industry over the next few years. While full elastic RTM is still some way off within the industry, Acceleware is well on the way towards taking the first step, by including a full elastic modeling capability in the AxRTM family of products. The first version of the elastic modeling product is scheduled for release later this year.
The second new AxRTM derivative product on the roadmap is Full Waveform Inversion (FWI). This is pretty much the Holy Grail in the exploration seismic industry. The ultimate aim is to have a computational tool which takes only recorded seismic data as input and automatically generates a detailed and accurate model of the Earth. This is very different to the current workflow, which requires a lot of manual input (e.g. from geophysical interpreters) over a long iterative cycle, usually involving many different computational tools, and in the end only results in an image of limited quality. With FWI it is imperative to have a high-performance wave propagation engine, such as that in AxRTM. Research on FWI is already at an advanced stage within Acceleware and is showing some great results. The current roadmap envisages release of a first FWI product in 2013. Interest from potential customers is already intense! Well, that’s all I have for now on this fascinating subject – I hope you find it useful.