What is Implicit Modelling?

Implicit modelling is a game-changing innovation in geological modelling.

Traditionally, geological models are produced using a manual drawing process. Sections are defined, and lithologies, faults and veins are drawn on the sections. Lines are then drawn to connect surfaces across multiple sections. Modelling geology in this manner is time-consuming and inflexible as it is difficult to update the model when more data becomes available. Early assumptions that later are proved incorrect could shape a model in a way that is never corrected because of the effort involved in starting over. Instead of using their knowledge to reveal important information about the study site, geologists spend significant proportions of their time engaged in mechanical drawing.

Implicit modelling, on the other hand, allows geologists to spend more time thinking about the geology. Implicit modelling eliminates the laborious legwork by using mathematical tools to derive the model from the data. A mathematical construct is built that can be used to visualise different aspects of the data in 3D. Leapfrog Geo uses FastRBF™, a mathematical algorithm developed from radial basis functions. FastRBF uses the data and parameters supplied by the geologist to derive any one of a number of variables to be modelled. Discrete variables such as lithologies can be used to construct surfaces, as well as continuous variables such as ore grades.

Instead of presenting a model constructed from rigid geometric constructs, the visualisations echo the natural forms found in reality.

What are the Advantages of Implicit Modelling?

Implicit models are easy to keep up-to-date with the latest data. New drillhole data can quickly be integrated, instead of taking weeks or longer manually modifying the model.

Implicit modelling allows several alternative hypothetical models to be produced from the data, quickly and easily. New data that affects the model, even in very significant or fundamental ways, can be assimilated and integrated with little effort. Models can be built rapidly, which means that a range of geological interpretations can be continually tested.

Because less effort is involved in creating a model, more time is available to spend on understanding the geology and studying more complex details such as faulting, stratigraphic sequences, trends and veins. The model can be developed to reflect reality to a greater degree of precision than was previously possible.

Geological risk is reduced when modelling is done implicitly. With traditional modelling, the effort involved means that the first model developed may be held to as the ground truth, despite mounting evidence that may discredit it. Instead, implicit modelling supports an approach that follows the proven scientific method, developing hypothetical models, experimenting to find new data to corroborate or discredit models and, ultimately, allowing the best model to be revealed. A geologist can experiment with alternative parameters at the limits of what is geologically reasonable to determine if there is any significant variation in the resulting models, which can then bracket the model with conceptual error bars. Geostatistical analysis can be conducted to identify what models are the most valid.

It is easy to change your mind when modelling implicitly. Perhaps models have been produced demonstrating isometric shells enclosing specific grades of ore. Commodity price changes then make it desirable to recreate the model using alternative ore grade values. If the model had been produced manually, doing so would not be practical. But with implicit modelling, the ease of generating a new model with new interpolation parameters means this valuable business information can be readily produced.

New questions can be answered more easily. A model tends to answer one question or class of questions well, and new questions require new models. If a model takes months to produce, those new questions may remain forever unanswered. If it can be produced with only days or hours worth of effort, valuable insights can be gained that could provide critical business value.

Implicit Modelling Makes Assumptions Explicit

Often, there is insufficient information in the data alone. For instance, drillhole data may well need to be supplemented with known details about the geology. When a geologist is constructing a model using traditional techniques, they use their knowledge of the geology to make decisions about the model as it is constructed. This is something the geologist will do automatically, which irretrievably conflates measured data with hypothesised data, hiding away subjective assumptions that have influenced the development of the model.

Implicit modelling, however, keeps the measured data separate from interpretations. The geologist can use polylines and structural disks to interpret the data without equating them to measured data. Implicit modelling makes assumptions explicit; there is a clear separation between hard data and user-introduced interpretations.

Implicit modelling and the presentation of a selection of models communicating different aspects of the geology provide new tools a geologist can use to communicate with professionals in other parts of the business.

On a purely business level, specialist staff can be putting their skills to use in productive, valuable geological modelling, rather than drawing lines on sections ad infinitum.

Implicit models are more repeatable and, therefore, more auditable, because they are derived from actual data and explicitly communicated geological interpretations, with selected parametric variables as inputs and processed using a mathematical algorithm.

The only thing implied in implicit modelling is the unknown value between two known values. Everything else is explicit. For this reason, it is best to refer to traditional modelling techniques as ‘traditional modelling’ rather than ‘explicit modelling’, assuming that it should be labelled with a name that is the inverse of ‘implicit modelling’. Implicit modelling is much more explicit than traditional modelling.

Best Practices

• Analyse data. Analyse your data using drillhole interpretation and data visualisation tools. Use 3D visualisation to look for errors in the data set.
• Stay focussed. Produce a model that answers a specific question or addresses a specific problem. Don’t unnecessarily model all the data available just because it is there. When a new question is asked, produce a model that answers that question using the necessary data.
• Experiment and explore. Produce variations of the same model, or even models, using quite different fundamental assumptions. Plan drillholes that will help reveal what model fits best and then discard models that are inconsistent with new data.
• Understand risk. Model using a range of input parameters and assumptions to understand the level of geologic risk.
• Share. Discuss and explore alternatives.
• Adapt. Previously, the effort of production and review of traditional models meant that there is reluctance to rebuild a model when new data becomes available soon after model completion. However, with implicit modelling, you should integrate new data and refine the model as soon as the new data is available. The revised model could indicate that planned activities should be redirected as expensive resources would be wasted persisting with the original plan, for little return.
• Evaluate and review. Don’t assume that because it’s easy to generate a model that you have quickly produced the right model. Understanding the geology is vital for validating the model and producing something that is geologically reasonable.

 

Got a question? Visit the Seequent forums or Seequent support