Most hydrocarbon reservoirs and the overburden rocks are faulted to some degree. Geologists often build fault framework models of key faults in a field, usually from seismic data. To make sure that our framework model is geologically credible, fault length and throw require QC - as there will be implications for wellplanning/drilling and for fluid communication if it is used in a dynamic simulation model.
Fault Length
Commonly, faults are interpreted (from seismic) as single continuous surfaces, when in reality they often form shorter segments separated by gaps or relay ramps (Figure 1). This is because seismic resolution does not in general allow for definition of faults located less than a few 100 m apart (Hesthammer & Fossen, 2000). Post drill, it may come as a surprise that wells on either side of a mapped fault are in pressure communication with each other, i.e., there is more than 1 fault in reality.
Fault length data can be collected from outcrop analogues (Figure 1).
Figure 1. Short faults separated by relay ramps (soft-linked). Kilve Pill, Somerset, U.K.
Fault Throw/Length
A good fault framework QC is to plot fault length against fault throw – the idea is that long faults should have bigger throws than shorter faults. Most geo-modelling softwares such as PETREL allow you to do this. A typical range for sedimentary rocks is 1:10 to 1:30 (Schultz et al. 2006).
The other problem with seismic data is that only fault throws greater than 15 - 20 m can be resolved (Fossen, 2010 and Figure 2A, 2B here). This is important as the fault tip (hashed zone in Figure 2A and 2C) can be under or overestimated. Ahead of the tip is a process zone (Figure 2A) of fractures; weak zones which are prone to mud invasion.
Clearly, we should try to avoid drilling through these tip regions. To minimise this risk, we should blank out the parts of the mapped faults where the throw is close to or below the expected resolution (i.e., hashed zone in Figure 2A). Then the fault can be continued based upon our (regional) knowledge of the throw gradients down to zero throw. A judgement then needs to be made on the extent of the process zone.
Figure 2 (A) Looking down on half of a fault polygon (white) with tip zone below seismic resolution (hashed) and surrounding damage zone on sides and process zone at right hand side (blue with yellow lines). modified from Fossen (2010, Figure 8.24). B) Cross-section through a - a' and (C) through b - b '.
Concluding Remarks
Faults are commonly more segmented and smaller than those represented on maps or in framework models. Getting the fault lengths right is important for history matching and prediction of communication pathways in the reservoir. Pin-pointing the fault tip also has value in the world of geomechanics as the rocks here are often weak, prone to mud invasion and should be avoided during drilling. We can use fault length/throw ratios and expected throw gradients to QC and correct our models.
References
Fossen, H. 2010. Structural Geology. Cambridge University Press.
Hesthammer, J, Fossen, H. 2000. Uncertainties associated with fault sealing analysis. Petroleum Geoscience, 6, 2000, 37 – 45.
Schultz, R.A., Okubo, C.H., & Wilkins, S.J. 2006. Displacement-length scaling relations for faults on the terrestrial planets. Journal of Structural Geology, 28, 2182-2193.