HELIOS SOFTWARE

Helios is the software implementation of the Wave Equation Based (WEB-AVO) seismic inversion technology, as traded by Delft Inversion.

HELIOS runs on the X86-64 based hardware or hypervisor with LINUX  or Windows based OS and uses an internet connection with HTTPS protocol over port 443 to connected to the Delft Inversion license server (license.delft-inversion.com). HELIOS consists of two parts: the main application Interface and the calculation nodes  (DInode).

Main Application Specification:

• 64 Bit LINUX: RHEL 8 / CentOS 8 / Debian 9 / Ubuntu 20.04; Gnome 3.13 or later, KDE 5.0 or later

• 64 Bit Windows: Windows 10 / 11; Windows Server 2019
• Docker 20.10 or later 

• CPU: Single CPU (4 or 8 core) with SSE4; Multiple CPU is preferred; AVX support is desirable; High clock speed and high cache is desirable

• Hard Drive Fast speed HDD (10-15K RPM); PCIe based SSD preferred

• Memory: Minimum 8GB; Preferred 32GB

• Graphics: Dedicated graphics card (not onboard); No benefit from high specification

Calculation Node (DInode) Specification:

We have provided a single node specification; this needs to be scaled to the project size and timelines. Please contact us for further information.

• 64 Bit LINUX: RHEL 8 / CentOS 8 / Debian 9 / Ubuntu 20.04 
• Docker 20.10 or later

• CPU: Single CPU (4 or 8 core) with SSE4; Multiple CPU is preferred; AVX support is desirable; High clock speed and high cache is desirable

• Hard Drive: Fast NAS storage
• Memory: 4GB Ram per core

Minimum:

• Seismic: PSTM or PSDM (ideal) offset gathers (time or depth) [SEG-Y]. Workflows also exist for angle gather and angle stack inputs.
• Velocity: Interval velocity from migration (time of depth) [SEG-Y]
• Well data: Sonic (DT), shear sonic (DTS), density, gamma-ray, resistivity, well track
• Interpretation: Horizons in TWT or Depth marking the top of the interval and all significant interfaces within the inversion window along with the respective well markers [ASCII]

Optional:

• Caliper [LAS],  gamma-ray [LAS] and well markers [ASCII]
• Lithology logs and facies definition [ASCII]
• Petrophysics Volume logs: Vclay, Vshale, saturation and porosity [LAS]
• Time-depth relationship
• Well completion /abandonment reports [PDF]

The run-time for WEB-AVO inversion depends mainly on the length and sampling of the inversion domain and the order of scattering requested to be modelled. For field data the run-time is approximately 5-10 times larger than what can be expected from conventional linear inversion. A target interval of 500-1000m length, vertically sampled on a 3-5m depth grid, results in run-times of 5-60 seconds per location, depending on available clock speed and memory of the available architecture. Given that all seismic locations are inverted independently, Helios is highly suitable for modern HPC architectures. Deployment of cloud computation further allows the user to optimise run-time according to his own time and budget constraints.

There is no easy way to answer this question because seismic processing has so many variables. The good news is that over the last years WEB-AVO inversion has been successfully applied to conventionally processed seismic pre-stack data with several thousand square kilometers surface coverage. Experience has shown that results can be improved by optimising the post-migration pre-conditioning workflow, specifically with respect to Radon demultiple, AVO scaling and time variant amplitude balancing. In an ideal scenario the seismic data would contain all elastic effects (interbed multiples, mode conversion, transmission effects) which are generated over the inversion interval with all overburden and surface related effects being attenuated or removed. We are more than keen to discuss your specific project challenges and to support the definition of an optimal processing sequence. Be invited to reach out to our team. 

Current implementations of WEB-AVO inversions in Helios (3D, 4D, PP+PS) are fully deterministic meaning that similar to conventional AVO inversion a low-frequency model is required as input. A major benefit of elastic wave-equation based inversion over linear techniques is that by deploying the non-linear relationship between the seismic data and the subsurface properties, inversion results can be obtained with a broader spatial bandwidth than what can be expected from the temporal bandwidth of the seismic input data. By extending the bandwidth of the inversion not only at the high-end but also at the low-end of the spectrum, the low-frequency model for WEB-AVO inversion serves as a starting model rather than a fixed low-frequency representation of the final inversion result. To what agree this bandwidth extension can be achieved from inversion of seismic field data depends on data quality but we have frequently seen that with the seismic spectrum starting at 7-8Hz, reservoir properties can be updated down to 3-5Hz. This unique feature can only be obtained from non-linear inversion techniques while linear inversion inevitably will have a one-to-one mapping of the seismic bandwidth onto the bandwidth of the inverted reservoir properties.

Quite a number of journal papers and conference abstracts have been published by our team, project partners and clients. A list of publications can be found on this website (www.delft-inversion.com/publications). You may also want to search for Delft Inversion on Youtube in order to watch some relevant videos covering technical and commercial information about our company and the offered technology portfolio. 

Below you can also find two EAGE E-Lectures related to WEB-AVO inversion:

By loading the video, you agree to YouTube's privacy policy.
Learn more

Load video

Always unblock YouTube

By loading the video, you agree to YouTube's privacy policy.
Learn more

Load video

Always unblock YouTube

While conventional linear inversion is routinely formulated  in terms of impedances (AI,SI), velocity ratio (vp/vs) or lambda-rho/mu-rho, the underlying elastic parameters of WEB-AVO inversion are the compressibility (inverse of bulk modulus) and shear compliance (inverse of shear modulus). This parameterisation is not purely based on the choice of its developers within Delft Inversion but rather follows from the fact that the elastic wave-equation can be formulated in these terms. Apart from being truly physically meaningful rock properties it turns out that these so called compliances have some very beneficial characteristics when it comes to quantitative characterisation of the subsurface. Compliances exhibit a natural orthogonality with the compressibility being most sensitive to porosity and pore fluid and shear compliance being mainly controlled by lithology. It can be demonstrated by a simple rock-physics analysis and or modelling that compressibility is three times more sensitive to hydrocarbons and other pore fluids than the acoustic impedance.  

We have a Helios Software Terms and Condition which is equivalent to the EULA and can be found here.