In his previous blog, Ian explained our decision to work with XGSLab, on the basis that it is an equally capable tool as CDEGS, that is much easier to use. In this blog, I’m going under to hood, so to speak. Comparing CDEGS with XGSLab to show you some real-world examples where CDEGS and XGSLab give almost identical results.
Therefore, in all of these examples, we have completed and verified the earthing design for these sites and found good agreement between the modelled results and the post-installation measurements. On this basis, we can say that XGSLab matches CDEGS for accuracy, and both tools have good agreement with the real-world measurements.
Comparing CDEGS with XGSLab – General Introduction
As a result, the 3 examples presented, we are comparing results from CDEGS HIFREQ against results from XGSA_FD. These are the high-end packages, offering the best agreement with the real world, particularly in low soil resistivity situations, such as those in example 2. So, for each example, we give an outline of the project, and then present the EPR at the point of fault. Together with, the EPR contour plot provided by each tool.
And, for consistency, the same soil models are used for both tools, generated using CDEGS RESAP. A future blog will look at the comparison between CDEGS RESAP and XGS SRA (Soil Resistivity Analyser)
Example 1
This is a wind turbine project, located in the countryside near Cardigan in Wales. The Wind Turbine connects to the 11 kV distribution network via a wood pole line with a small package substation to the bottom left. A transformer located at the base of the turbine feeds a Farmhouse (not included in the model. A key requirement on this project was to keep the hot contours away from the Farmhouse and other 3rd parties.
Soil Model
CDEGS HIFREQ
Another note is that the first phase of the model, where CDEGS determines the current distribution (and therefore the voltages) in the conductors, takes 10 minutes to run on my PC. The second phase, where it determines the voltages at the soil service takes 15 minutes
XGSA_FD
XGSA_FD takes 1 minute 22 seconds to perform the first phase of modelling and 40 minutes to run the 2nd phase
Comparison
| Item | CDEGS HIFREQ EPR (V) | XGSA_FD EPR (V) | Difference (%) |
| Value | 3560 | 3561 | 0.03 |
Example 2
This is a 33 kV substation that was constructed as the point of connection for 2 large scale solar farms near Bristol.
Soil Model
CDEGS HIFREQ
XGSA_FD
Comparison
| Item | CDEGS HIFREQ EPR (V) | XGSA_FD EPR (V) | Difference (%) |
| Value | 286.8 | 287.6 | 0.27 |
Example 3
This is energy from a waste site located near London. There are 2 generators located on-site, each with a dedicated grid connection, with the point of connection on an isolated earth electrode. This model simulates an earth fault in the larger building.
Soil Model
CDEGS HIFREQ
XGSA_FD
Comparison
| Item | CDEGS HIFREQ EPR (V) | XGSA_FD EPR (V) | Difference (%) |
| Value | 41.42 | 41.40 | 0.06 |
Conclusions
In these example cases, we’ve shown that in all cases, XGSLab and CDEGS have an extremely good agreement. The team at SINT Ingegneria have produced papers showing that XGSLab produces results matching the standard examples in IEEE 80.
So the difference we found between XGSLab and CDEGS
| Item | CDEGS HIFREQ EPR (V) | XGSA_FD EPR (V) | Difference (%) |
| Example 1 | 3560 | 3561 | 0.03 |
| Example 2 | 286.8 | 287.6 | 0.27 |
| Example 3 | 41.42 | 41.40 | 0.06 |
Comparing XGSLab and CDEGS, a key point of difference is in conductor segmentation. The default CDEGS segmentation can lead to poor accuracy, particularly in low resistivity soil structures like those in example 2. In comparison, XGSLab makes assigning segmentation for the whole model a 1 click process, and the default segmentation option leads to much better accuracy. Obviously, any increase in segmentation level is likely to lead to an increase in computation time. However, getting the accuracy right is more important, and in my experience, a few minutes increase in computation time is not as important as getting things right the first time. It is likely that the increased difference in example 2 can be put down to the difference in segmentation approaches
XGSLab UK Agents – GreyMatters
Now we’ve shown that for comparable models XGSLab and CDEGS will give the same results. So, why not get in touch to request a demo and see for yourself how easy it is to use XGSLab.
