Error: "RBF Interpolation failed to converge. Check the accuracy. If using a Spheroidal interpolant, a smaller range or larger nugget value may help".

Occurrence:

While running a Structured Trend or Numeric Model I receive the following error:

RBF Interpolation failed to converge. Check the accuracy. If using a Spheroidal interpolant, a smaller range or larger nugget value may help

Resolution:

1. Check the input data to see if there are any duplicate or almost duplicate input points/intervals, as well as any outliers
2. Delete these if possible.

For example, it may be that one of the drillholes added to the dataset is either a duplicate of an existing hole, in the wrong place, or has the wrong set of codes on it, all of which could cause issues with the RBF interpolation.

Root Cause:

Generally, duplicate or redundant input data must be removed.  

Notes:

When Leapfrog creates surfaces using the RBF fitting algorithm, the software goes through different iterations of the surface until it finds an acceptable solution for it that fits the input data.  If Leapfrog hits the iteration limit without finding an acceptable surface, then this error is displayed. (A limit is required to avoid processing infinitely without reaching a solution). 

Commonly this is related to the density of the input data e.g. very dense input dataset with a lot of points/intervals either very close or actually duplicates of each other, especially when some of the data is contradictory.  For example, this can often happen when drillholes cross over each other a lot in the input data resulting in very closely spaced data points.