Grounding Analysis for Utility Scale Photovoltaic Power Plant

Utility scale systems (5 MW or greater) present several challenges for properly designing grounding system for personnel protection concerns. This discussion, given by David Lewis, PE, Grounding and Power Systems at EasyPower, highlights some of these challenges and provide methodologies to accurately assess the grounding system performance with regard to IEEE Std 80.

Download a Free Demo

David's Presentation (PDF)

Thumbnail to download the PDF of the presentation
Download Presentation

Answers to Q&A Following Live Webinar

Question Answer (David Lewis, PE at EasyPower)
Is XGSLab able to work with Imperial Units? XGSLab is used internationally, and users can perform their work in imperial or in metric units.
Can you tell us what is the simulation time duration for such a solar farm with all the steel supports modeled please? The XGSLab software computes the metallic potential and then the user can select the area calculation for touch/step evaluations. For the presented case studies, the first stage completed in 20-30 minutes and the area calculations took 20-30 minutes. The latest version of XGSLab performed the same area calculation in 2 minutes. Note that simulation time is dependent upon computer configuration, system size, and step size, and for this presentation the case studies performs the calculation an i5, 6 GB laptop.
Typically the DC side of the inverter is isolated from ground. The AC side of the inverter is also connected with a cable to the transformer. In that case there is no ground fault current through the soil which creates STEP & TOUCH potentials and GPR. Therefore do you really need such a large ground grid for the PV station? The faults that are typically evaluated in the utility scale PV site, like that of the presentation, are located on the medium voltage (the 12-34 kV) at the inverter-GSU. A local  fault on the LV side of the GSU would not produce a GPR as they have a metallic connection back to the source, as you noted. One of the challenges of PV grounding design is understanding how the system under analysis is actually connected, as there are different configurations. In many utility scale PV systems the grounding system is common from the DC grounding conductors and the AC grounding conductors.
Is this grounding program an add on to Easy power? Or a completely separate program? XGSLab is a separate software from the Easypower software. It is  specifically designed to analyze grounding, lightning, and electromagnetic interreference studies. Please let us know if you would like to learn more, and feel free to visit https://www.easypower.com/products/xgslab-grounding
Can you remind me from what level, the step and touch voltages is of concern, in a installation. Based on IEEE Std 80, the Step and Touch Voltage permissible voltage is dependent on the size of the person, soil characteristic, consideration of surfacing material, and the fault clearing time.
Should we bond the fence grounding grip to the PV plant grid? It depends on the scenario, but if your perimeter fence was not exposed to the utility sourced fault you may be transferring hazardous voltage to the perimeter fence when directly connected to the PV plant grid.  
Can steel conduits being considered in the model? Yes you can consider steel conduits in the model, and if you can even model the faulted conductors path to incorporate the mutual impedance on the conduit.
Is there any difference in grounding characteristics if concrete footing is chosen over piling for the array posts? Concrete often has a similar resistivity to the native soil resistivity characteristic, so modeling the concrete  may not yield different results compared to the auxiliary ground electrode. That said, this is dependent on other factors that drive soil resistivity. 
Is it possible to modelling some elements without any connection to the main grid? Such as an independent fencing around the facility. The XGSLab software can model several intendent grounding systems, and this is often required for evaluating voltages transferred from an energized grounding system through the soil to adjacent metallic objects.
How do you model the poles of the trackers? The poles of the trackers are modeled as steel conductor extending into the earth to match the designed buried depth of the support posts. I was using the GSA_FD module which limits modeling below grade, so modeled my PV support track connection as insulated conductor. With the XGSA_FD the engineer could model the system above and below the earth surface.
How could we detect the impact of the DC side in PV systems, considering that both sides DC/AC can be connected on the same Grounding System. In the XGSLab software, the engineer could model the grounding system as connected in the design to the evaluate the total grounding system performance. The end of the webinar shows an example of using the PV posts to reduce the GPR, and improve touch/step voltage results.
Regarding the modeling of the support posts do you do if there are too many elements. Do you reduce the number? Yes, if you encounter issues with the number of elements you can model less of the posts and anticipate a more conservative result. The engineer performing these studies should understand the impact of including or excluding the auxiliary grounding components and proceed with best engineering judgement.
What is the greater distance we can consider an equipotential model? This is dependent upon the soil resistivity and density of the grounding system. The XGSLab User Guide includes a 'Application domain of GSA and GSA_FD that gives a chart for the application limits of an equipotential model.
Should we consider the seasonal variation of soil resistivity? This topic could be its own webinar, but grounding analysis should consider seasonal variations that provide more conservative results. Significant increase in soil resistivity occurs during frozen conditions, but how much this affects an analysis is dependent upon the grounding electrodes depth and freezing depth. For example, shallow freezing may not affect the grid resistance and acts as an insulating layer. Depending on when the measurements were performed, you may need to consider lower resistivity from warmer soil. XGSLab will be incorporating a new tool in the next version of the software to help users navigate seasonal affects on soil.
Can the plot be imported from a CAD dwg file? Yes, XGSLab can import a dxf to quickly develop the model. It can also export the system to a dxf and export the GPR, Touch/Step contours to a dxf to aid users in evaluating their systems performance, and adjusting designs as determined by the study.
Can you avoid having to put in a ground grid if your support posts are driven deep enough? We show in the case study that the grounding conductor may be reduced considering auxiliary grounding systems like posts, but a ground system is still required.
Is the step voltage still an issue if you have insulated shoes? Considering shoe resistance increases the permissible step voltage that personnel may withstand. Often a grounding study does not include the shoe resistance, considering areas that may be accessible to the public like the exterior of the fence. An engineer could consider shoe resistance in controlled area, but consider environmental or work practices that could significantly reduce the shoe resistance.
Does the program simulate DC currents? XGSLab can simulate DC systems and up to 100 MHz.
What injection fault current are you using in this model? The examples shown have faults injected at the end of the MV lines, where it connects to the inverter-GSU, and at the collector substation is an energized conductor removing the local delta contribution.
Can this software can be use for any type of facility? The software can be used to evaluate grounding system performance for substations, industrial sites, generation facilities, and many other systems. If you have a specific application that you are thinking of, please let me know and I can help clarify for your application.
How will you know you have a grounding problem? Regarding grounding analysis for personnel/public safety, an engineer can use the XGSLab software to evaluate touch and step voltage hazards. Please check out our Grounding 101 video https://www.easypower.com/products/xgslab-grounding or give us a call to discuss.
Are there other analysis approaches that do not require the PV posts to be modeled? There are other valid engineering solutions to performing these types of studies. For example, modeling lumped impedances of parts of the grounding system or just modeling the grounding conductor. With any approach the engineer should understand the impact of their process and proceed with best engineering judgement.