Substation ground grid — IEEE 80 step / touch voltage design

Reference notes

A substation ground grid (earthing grid, earth mat) is a mesh of buried bare-copper conductors and vertical ground rods installed under and around every substation. Its primary purpose is personnel protection from step and touch voltages during ground faults. Design governed by IEEE Standard 80 (Guide for Safety in AC Substation Grounding, latest revision 2013 with corrigenda 2015).

Three functions of the ground grid

1. Provide a low-impedance return path for ground-fault currents so protective relays (51G/87G) operate reliably.
2. Safely dissipate fault current into earth without damaging equipment or causing fires.
3. Limit STEP and TOUCH voltages below tolerable shock thresholds to protect personnel. This is the most rigorous design constraint.

Step vs touch voltage

• STEP voltage: potential difference between two points 1 m apart on the earth's surface (operator's feet positioned one in front of the other). Current path: foot-to-foot through lower body.
• TOUCH voltage: potential difference between hand (touching a grounded metal structure) and feet (standing on earth nearby). Current path: hand-to-feet through chest and heart — the worst path for ventricular fibrillation.
• Without proper grid design, both can reach KILOVOLTS during a substation ground fault.

Tolerable shock voltages — Dalziel body-current threshold

• Body-current limit at fibrillation threshold:
  • 50-kg (110-lb) person (general public access): I_body = 0.116 / sqrt(t_s) amps.
  • 70-kg (155-lb) person (trained substation personnel only): I_body = 0.157 / sqrt(t_s) amps.
• Where t_s is shock duration in seconds (typically equal to fault clearing time).

• ρ_s = surface-layer resistivity (typically gravel at 1000-5000 Ω·m — much higher than native soil).
• C_s = surface-layer derating factor (≈ 1 for thick gravel, 0.6-0.8 for thin).
• For ρ_s = 3000 Ω·m gravel, C_s = 0.8, t_s = 0.5 s: V_step ≈ 2400 V, V_touch ≈ 740 V.
• Without surface gravel: V_step might be only 130 V, V_touch only 80 V — essentially impossible to design to. Surface gravel is critical.

Ground potential rise (GPR)

• GPR = I_g × R_g — the voltage of the ground grid relative to remote earth during a fault.
• I_g = ground-grid current = I_fault × S_f × D_f.
  • S_f = current-division (split) factor — typically 0.5-0.75 for transmission stations with multiple parallel ground paths.
  • D_f = decrement factor accounting for DC offset and X/R ratio — typically 1.0-1.65.
• R_g = grid-to-earth resistance. Sverak formula: R_g ≈ ρ × (1/L_total + 1/sqrt(20·A)) + depth term.
• Typical substation GPR during a fault: 500-5000 V.

Soil resistivity (IEEE 81)

• Wenner 4-probe method: four equally-spaced probes; inject current through outer two; measure voltage between inner two. Apparent resistivity: ρ = 2·π·a·V/I where a = probe spacing.
• Repeat at multiple spacings (1, 2, 5, 10, 20, 50, 100 m) to derive a 2-layer (upper ρ_1 of thickness h, lower ρ_2) or 3-layer soil model. Depth probed ≈ probe spacing.
• Typical values:
  • Dry sand: 1000-10,000 Ω·m
  • Clay: 5-100 Ω·m
  • Rock: 1000-100,000 Ω·m
  • Sea water: 0.1-1 Ω·m
  • Typical substation site: 100-500 Ω·m
• Seasonal effects (frozen ground, dry summer) matter — measure in worst-case condition.

Mesh & step voltage design — Sverak formulas

• V_mesh = maximum touch voltage anywhere within a mesh of the grid.
• K_m = mesh geometric factor — depends on conductor spacing D, burial depth h, mesh divisions, ground-rod presence. Tighter spacing and rods reduce K_m.
• K_s = step geometric factor — typically smaller than K_m at grid interior but larger at corners.
• K_i = irregularity factor — non-uniform current distribution; higher at corners.
• L_M = effective conductor length (includes weighting for rods).

Pass criteria

• V_mesh < V_touch_tolerable (so touch voltage anywhere in the grid is safe).
• V_step < V_step_tolerable (so step voltage at the perimeter is safe).
• If FAIL: tighten mesh, add ground rods, deepen burial, increase surface gravel resistivity, deep wells with low-resistivity bentonite, or chemical ground enhancement.

Design iteration

1. Estimate I_fault, S_f, D_f → I_g; estimate t_s from protection clearing time.
2. Measure soil resistivity (Wenner 4-probe, IEEE 81); fit 2-layer model.
3. Pick tentative grid geometry — area, conductor spacing, burial depth, rod count and placement.
4. Compute R_g, V_mesh, V_step from Sverak formulas (modern software: CDEGS, ETAP GroundMat, SKM GroundMat).
5. Compare V_mesh vs V_touch_tolerable and V_step vs V_step_tolerable.
6. If any fails, iterate geometry or add mitigation.

Practical grid layout

• Rectangular mesh of bare-copper conductors typically 4/0 AWG to 500 MCM (kcmil), buried 0.5-1.5 m deep.
• Mesh spacing: 3-10 m in main yard; 1-3 m near operator workplaces (panel rooms, breaker bay access).
• Ground rods (10-ft / 3-m copper-clad steel) at every fence post, corner, and along perimeter. Rods reach lower-resistivity strata if surface soil is high-resistivity.
• Equipment grounding: every transformer tank, switchgear cubicle, lightning arrester base, fence post, and structural steel bonded to the grid.

Fence grounding

• Fence bonded to grid (most common modern practice): fence is at full GPR during a fault. Install an outer perimeter ground ring 1 m outside the fence and 1 m deep — raises the earth potential immediately outside the fence to nearly the fence potential, eliminating touch-voltage hazard from outside.
• Fence isolated from grid (less common): separate fence-ground rod; gradient at fence base controlled to < V_touch_tolerable. Requires careful isolation of gates, ladders, signs.

Transferred potential — a critical hazard

• Communications cables, water pipes, signal wires entering the substation from REMOTE EARTH can be at remote-earth potential.
• Touching them inside the substation while standing on the grid → full GPR shock — potentially lethal.
• Mitigation:
  • Isolation transformers on telecom circuits.
  • Fiber-optic for SCADA / protection signaling (galvanic isolation).
  • Neutralizing transformers on conductive telecom.
  • Insulated couplings on water/gas pipes at the substation boundary.

Validation — IEEE 81 fall-of-potential test

• After construction, measure R_g by the FALL-OF-POTENTIAL test per IEEE 81:
  1. Drive current probe 250-500 m from substation (≥ 5× the largest grid dimension).
  2. Drive potential probe progressively closer along the line between substation and current probe.
  3. Inject test current; measure voltage between grid and potential probe.
  4. Plot V vs distance. Flat region = true R_g. Tagg's 61.8% rule for homogeneous soil.
• Acceptable: measured R_g within 10-20% of design value.
• Spot-check V_step and V_touch by injecting fault-replica current and measuring surface voltages with high-Z voltmeter at critical locations.
• Documentation per NETA ATS (Acceptance Testing Specification); periodic re-test per NETA MTS, typical 5-year interval.

Problem sites

• High-resistivity rock or permafrost installations may not achieve target R_g.
• Mitigations:
  • Extended counterpoise — bare conductors extending 100 m out from substation.
  • Deep wells filled with low-resistivity bentonite or chemical compounds.
  • Chemical ground enhancement (CHEMrods, Erico Cadweld bonded).