Grading approach selection is one of the largest variable cost decisions in utility-scale solar. On a single project with challenging rocky terrain, three approaches produced earthwork costs ranging from $335K to $1.06M. Same site. Same panels. Same capacity. The only variable was the grading method.
This article walks through all three approaches with real data from a PVX engineering analysis.
The Site
The project site presented extreme terrain difficulty:
- 44% of the area classified as very hard rock (concrete/asphalt-grade)
- 3.5% classified as hard (limestone/gravel)
- Slopes reaching 40-45% in localized areas
- A mix of medium-soft, firm, and hard soil across the remaining area
Before any grading strategy was selected, PVX.Cad mapped the slope distribution (north-south and east-west) and classified soil hardness across the full site envelope. This terrain intelligence informed every decision that followed.
Approach 1: Full Terrain Smoothing
The conventional method. PVX.Cad’s smoothing algorithm was applied to the digital terrain model, targeting a maximum 20% slope across the site. The algorithm balanced cut and fill, shaving hilltops and filling depressions to create a flatter surface.
Results:
- Total cut: 118,225 m3
- Total fill: 102,883 m3
- Net excess: ~15,343 m3 (requiring off-site removal)
- Cost: $1,062,481
Cross-section verification showed elevation differences exceeding 2 m between the original and designed surfaces. The smoothing algorithm brought most of the site within tolerance, but several localized areas remained problematic. 3D verification in PVX.View confirmed that some rack positions still had legs buried into terrain due to residual irregularities the global algorithm could not resolve.
This approach moves the most earth, costs the most money, and still leaves edge cases unresolved.
Approach 2: Pile-Adaptive Local Grading
Instead of grading the entire site to a uniform surface, PVX.Cad’s “Adapt Terrain to Pile” function performed cut/fill operations only at each rack position. Each PV table was adapted to the existing terrain using a fixed rack base height of 0.60 m.
Results:
- Total cut: 48,109 m3
- Total fill: 10,844 m3
- Net excess: ~37,265 m3
- Cost: $438,046
The analysis confirmed that no rack or pile exceeded design limits. All leg lengths and slope angles remained within acceptable tolerances. The pile-adaptive approach achieved full constructability without the massive earthwork volumes of conventional smoothing.
A direct cross-section comparison at the same location showed the fundamental difference:
- Smoothing: 3.0 m cut depth
- Pile-adaptive: 0.8 m cut depth
The 3D model confirmed that PV tables followed the natural terrain topography without valley filling or hilltop excavation.
Savings vs Approach 1: $624,435
Approach 3: Table Splitting + Pile-Adaptive (Hybrid)
On undulating terrain, some locations showed surface irregularities where standard 2x26 table configurations produced pile lengths approaching or exceeding the 4 m structural limit. PVX.View’s 3D inspection tool flagged these critical zones.
The solution: split the affected tables from 2x26 to 2x13 configuration. 52 tables were selected for splitting. This did not change DC power or panel capacity. The number of columns increased, but each table became shorter, conforming better to local terrain contours.
Results:
- Total cut: 34,819 m3
- Total fill: 14,472 m3
- Net excess: ~20,347 m3
- Cost: $335,376
Cross-section comparisons confirmed that the longest rear pile dropped from ~5.06 m (undivided table) to ~3.93 m (split table), bringing all piles under the 4 m limit.
Savings vs Approach 1: $727,105 Savings vs Approach 2: $102,670
The Full Comparison
| Approach | Method | Total Cut (m3) | Total Fill (m3) | Total Cost (USD) |
|---|---|---|---|---|
| 1 | Full terrain smoothing | 118,225 | 102,883 | $1,062,481 |
| 2 | Pile-adaptive local grading | 48,109 | 10,844 | $438,046 |
| 3 | Table splitting + pile-adaptive | 34,819 | 14,472 | $335,376 |
Why the Cost Difference Is So Large
The financial model accounts for terrain reality, not theoretical averages:
- Excavation cost: $3.00/m3 (weighted by soil hardness distribution, higher for hard/very hard zones)
- Fill placement and compaction: $5.00/m3
- Hauling excess cut off-site: $4.00/m3
- Hard soil pile installation: $15/pile (drilling) to $25/pile (drilling + concrete)
- Indirect site costs: 10% of direct earthwork
With 48% of the site in hard or very hard soil categories, excavation costs are significantly higher than standard pricing. A tool that does not classify soil hardness before grading will underestimate costs every time.
What This Means in Practice
The difference between these three approaches is not theoretical. It is the difference between a project that hits budget and one that blows it.
Approach 1 is what happens when terrain is treated as a surface to flatten. Approach 3 is what happens when terrain is treated as a constraint to design around.
Both produce a buildable result. One costs $727K more.
The analysis was performed in a single design session inside AutoCAD using PVX.Cad, with 3D verification in PVX.View. All three scenarios generated from the same terrain model, compared side by side, and validated before any earthwork began.
Engineering analysis by Mustafa Unal. Full case study data available at pvx.ai/customers.