How To Improve The Predictability Of Your Reservoir Model With The Smallest Effort Possible (Part One)

To begin with, I think you agree with me:

Thanks to the progress in computing performance and commercial geomodelling software, building a reservoir model are not terribly difficult. But to build the ‘right’ model with adequate level of complexity that can satisfactorily support the business decision is often not straightforward, at a particular stage of field development.

Very true, we see that:

In many cases, we tend to waste a lot of time building detailed models with unnecessary complexity that is only deductive in available data. In contrast, sometimes the models we build fail to support the business decisions due to the insufficient realism and accuracy.

Sounds familiar? Sure, you may have been there too. 

The first scenario is referred to as “over Modelled”, the later one is referred as “under Modelled”. In the end of the day, either way, can lead to wrong business decisions, which degrade the project profitability.

Have you thought about this before? 

What on earth is the ‘right’ model to build, and how to handle it with fit-for-purpose complexities and information to satisfy the business need?

The bad news is that this rarely taught. Good news is we are talking about it now in this post,

I will give you some easy-to-follow tips on how to build the just ‘right’ model, with very few changes in your daily modelling routine.

Three principles that dictate the degree of complexity of a reservoir model:

  • The right modelling methodology keeps the reservoir model as simple as possible with which it can still address your key technical issues.
  • A modelling strategy is not fit-for-purpose until the aim of it is clear. Unnecessary complexities should always be avoided unless the additional information is critical to the problem.
  • In general, when not biased, matching more data means better insight into the reservoir facts with fewer subsurface uncertainties.

With these three principles in mind, your first takeaways are several useful understandings:

  • Complexity doesn’t guarantee you a better model
  • A highly detailed all-inclusive Model ≠ accurate model. It does more harm than good.
  • The key to increase model accuracy and reduce subsurface uncertainty is to incorporate the just data in need.

Take a look at this diagram by Our Reservoir modelling Advisor Raffik Lazar, this is something that is really interesting.

Type of modelling strategies (black color text) based on Understanding of the reservoir (X-axis), Amount of data available to build the reservoir model (Y-axis) and purpose of the reservoir model (red color text).

Back to what I was saying,

“Over Modelling” tends to occur towards the exploration and the early stage of the development lifecycle, while “Under Modelling” is more likely to take place towards the mid-and-late stage of the field life.

Here is the most important part.

It is relatively easy to address the “Over Modelling” issue. You just need to go back and review the purpose of your modeling task. Here are what you should consider for a lean and effective model:


It sounds simple. And it is.

If you can effectively embed the above into your modeling practice, you will be able to customize the modelling effort depending on the amount of data, the level of understanding and the purpose of the model, efficiently.

However, on the other hand, “Under Modeling” is a much more challenging issue. The ability to build reservoir models of high realism is the prerequisite for locating the remaining oil.

“At low oil price, this is particularly important given that the failure to reach the sweet spot may become a disaster to the economics of a well, sometimes even that of a small marginal field.”

However, standard modelling workflows often let us down. The models are usually large and unwieldy, the integration of production data is time-consuming and the incremental nature of the data accumulation means models tend to become ‘patched’. Models are commonly passed hand-to-hand between practitioners to the point that ownership is lost. The update and maintenance of the ‘field model’ becomes a job in itself, often separate from the process of managing the mature field. The modelling process thus reaches a technical limit, and loses its value (Oxlade, 2015).

In the next article, I will talk about “Under Modelling” problem in more details.We will show you some powerful tools to effectively integrate all the necessary data into your model. 

If you think this article is useful, Stay tuned and follow our company page iRes-Geo to read more articles about this topic.

Reference: “The concept of right modelling, toward a more efficient use of the static modelling tool” by Raffik Lazar

Author: Dr. Yi Huang

Founder of iRes-Geo,

Formerly Sr. Reservoir Geophysicist Statoil, 10 years + Experience in Integrated Reservoir Geophysics R&D and Technology Implementation

MSc. Reservoir Engineering, Ph.D. in Reservoir Geophysics, Institute of Petroleum Engineering, Heriot-Watt University, Edinburgh