Hoolock Consulting will be presenting a Tech Byte at the Seismic 2019 conference in Aberdeen in May. Go to https://www.spe-aberdeen.org/events/seismic2019/ for more information about the event. We will be presenting “Reduce risk and cost by using multi-attribute seismic analysis and machine learning”. This is our accepted abstract.
The generation of seismic attributes, such as amplitude, dip, frequency, phase and polarity, has enabled geoscientists to interpret more geological features in their seismic data. The combination of attributes, with the proper input parameters, can produce even more meaningful results and help reduce risk in prospects and projects.
Much work has been done to understand how to combine multiple attributes into volumes that can be used to define more geology and improve seismic interpretations. With increased computer power and machine learning, self-organizing maps (SOM), a form of unsupervised neural networks, has proven to be an excellent method to do this. SOM analysis works by clustering samples based on their attribute values and has been beneficial in defining stratigraphy, seismic facies, DHI features etc. Recent work utilizing SOM, along with principal component analysis (PCA), has revealed geologic features that previously were not identified or easily interpreted from the data.
The presentation will show how self-organising maps work and provide examples of how they have been used to develop a deeper understanding of the subsurface than can be seen on standard seismic displays.