Research Result Summaries

• 1: Substation H-Frame Inspection Report
• 2: Substation Ground Grid Corrosion Study 1
• 3: Substation Ground Grid Corrosion Study 2
• 4: Corrosion Monitoring System

1 - Substation H-Frame Inspection Report

EPRI was asked to provide a condition assessment of an H-Frame within a substation. The survey consisted of four techniques, the first SSPC VIS II to quantify the degradation of the coating syste and to understand when the structure would begin to lose structural strength. The second technique was “ISO 23277 Non-destructive testing of welds” using dye penetrant for inspection of weld joints to identify inclusions or crack propagation The third was “ASTM E797/E797M-21 Standard Practice for Measuring Thickness by Manual Ultrasonic Pulse-Echo Contact Method” to identify reduced wall thicknesses due to internal corrosion. The fourth was the use of “ASTM Standard G46” and pit gauges to quantify any external corrosion.

Survey Approach

All weld joints (W) and wall thickness (T) at critical locations were identified as inspection points on the structure. An inspection platform and lineman were provided for access to the structure and fall protection harnesses were used for a safe inspection. Additional inspection points are at clevis supports to understand the severity of fretting corrosion. Fretting wear is the mechanism causing mass loss due to conductor motion where flash rust is wiped from the surface and reforms with the presence of moisture.
Technical note: it is the utilities responsibility to determine acceptance criteria for all sectional losses which will be used as a benchmark during the survey.

Inspection Results

Direct Assessment:

1. The coating system integrity had been compromised and iron oxides had already formed (see Figure 3).
2. Fretting wear was noted at all insulator attachment points resulting in a significant elongation of the attachment point for the clevis (see Fretting Corrosion at Figures 2).

Structural Measurements

1. The design wall thickness for the vertical members was not available from archives, however the horizontal member wall thickness specification was .467 inches. The standard deviation for the vertical structural member was considered unremarkable and within manufacturers mill standard allowances. The horizontal member was also unremarkable for the standpoint of manufacturers mill allowances however this highlights the sensitivity of an ultrasonic inspection method.
2. The dye penetrant inspections did not reveal any weld cracks or inclusions however there are initiation points for monitoring.
3. The underside of the horizontal member revealed some localized (pitting) corrosion (see Figures 1,2 & 3 and Table 1)

Vertical Support		Horizontal
Inspection Points	Wall Thickness (inches)	Wall Thickness (inches)	Sectional Loss (%)
1	0.668	0.449	3.85%
2	0.656	0.448	4.07%
3	0.676	0.445	4.71%
4	0.670	0.449	3.85%
5	0.665	0.443	5.14%
6	0.667	0.443	5.14%
7	0.676
mean	0.669	0.447	4.4%
std dev	0.008	0.003	0.006

Environmental Data

Contaminant levels were found to be unremarkable however the humidity levels were elevated which increase time of wetness and resulting surface corrosion (see Figure 4).

Surface Contaminants	PPM
Chlorides	30
Nitrates	5
Sulfides	3

Figure 1 – area at base of shield wire support (East) – ASTM G46 (A5, B2, C2)

Figure 2 – Area at Insulator Drop – ASTM G46 (A4, B3, C2), See Note #2

Figure 3 – area at base of shield wire support – General corrosion

Figure 4 – underside of conductor support

Conclusions and Recommendations

EPRI structural surveys are constructed by quantifying coating system integrity and then characterization of any wear signs, cracks or sectional loss due to corrosion. The coating system was judged to be in poor condition, so this moved the focus of the survey to inspection points dictating structural integrity. Weld joints were evaluated for crack initiation and propagation, wall thicknesses were measured at locations considered to be stress risers or areas more prone to degradation and insulator attachment points were evaluated for wear. There were no visible signs of crack initiation and wall losses were estimated to be 5% at inspection points. Based upon these findings the structure should be considered in serviceable condition. It is recommended that engineering determine acceptable localized wall losses but the coating system should be repaired and placed in a 20 year inspection cycle.

2 - Substation Ground Grid Corrosion Study 1

Introduction

Gaps were identified within IEEE Std 81 “Guide for Measuring Earth Resistivity, Ground Impedance, and Earth Surface Potentials of a Grounding System” for identifying and locating corrosion damage within substation ground grids. To satisfy these gaps, EPRI has been developing methods based on soil characterization and electrochemistry to augment the existing methodology.

Objective

Soil characterization provides an understanding of soil corrosivity and how asset materials, such as copper, zinc (galvanizing), and steel, perform throughout long-term environmental exposure. Soil-exposed assets include the foundations for bus work, tank bottoms, and the ground grid itself. Results of soil characterization may be used in a host of applications that include site planning, forensics, and optimizing maintenance operations.

Background

EPRI was contacted to understand the corrosivity of soils located in and adjacent to a substation. A field crew completed a visual inspection of the facility during a walk-through, and then collected soil samples at various locations (Figure 1). These locations were selected based on the NRCS Database “SSURGO”, where the US Soil Survey’s maps are compiled, delineating soil series taxonomic groupings throughout the country. When a soil has similar characteristics as defined by the soil surveyors, it is mapped as a polygon and labeled as a “series”. Sample locations at the Manchester Substation were selected to be representative of the three native soil series adjacent to the substation assets. Each soil series was sampled with three replicates.

The collected soils were packaged, labeled, and shipped to the Charlotte Corrosion and Soils Laboratory for analysis. EPRI’s soil characterization consists of chemical analysis, particle size analysis, and linear polarization resistance using electrochemistry. The following metrics were measured during analysis:

• Chlorides
• Nitrates
• pH
• Calcium Chloride
• Humic Matter
• Cation Exchange Capacity
• Base Saturation Percent
• Exchangeable Acidity
• Calcium
• Magnesium
• Manganese
• Zinc
• Sodium
• % Clay
• % Sand
• % Silt
• Color (Hue, Value, Chroma)
• Corrosion Rates for zinc, steel, and copper as a function of moisture and temperature

Methodology

The following is a summary methodology of EPRI’s soil corrosion evaluation procedures. For an extensive overview, please reference EPRI PID 3002018899 “Recommended Practice Guidelines for Corrosion Inspection and Assessment (2020).”

Chemistry

pH testing is based on ASTM D4972-13: Standard Test Method for pH of Soils. The pH of a soil is a useful variable in determining the solubility of soil minerals and the mobility of ions in the soil.
Chloride and nitrate concentration measurements are based on Colorado Department of Transportation Laboratory Procedure (CP-L 2104): Determining the chloride ion content in water or water-soluble chloride ion content in soil and Texas Department of Transportation Procedure (Tex-620-J): Determining chloride and sulfate contents in soil. Chlorides and nitrates are extracted from the soil into a deionized water solution, which is then filtered to remove any remaining soil particulates. The actual measurement of the chloride and nitrate ion concentrations is completed using Ion Selective Electrodes (ISE) for each of the ion types.
The remaining chemical analyses are completed by the North Carolina Department of Agriculture due to the availability and expediency of testing. The full table of chemistry results from this testing, along with descriptions of calculated parameters, can be found in Appendix A.

Particle Size Analysis (PSA)

EPRI evaluates particle size using a coarse sieve stack and Beckman-Coulter laser diffraction for the fine fraction distribution (Figure 2).

Corrosion Rates

EPRI collects corrosion rate data using an electrochemical test method called MT5 to highlight a range of corrosive environmental conditions. The procedure was created to understand the impact of soil moisture and temperature on the corrosion rate of copper, steel, and zinc. In this procedure, a soil is subjected to varying moisture and temperature plateaus, then corrosion rates are measured for different metals. Electrical conductivity, pH, and native potential are also measured. Five combinations of temperature and moisture are measured in the MT5 procedure from Table 1: Cold Moist, Hot Moist, Room Wet, Cold Saturated, and Hot Saturated.

Table 1: MT5 Parameters.

Level	Range
Soil Temp.	Hot	104 - 111 (˚F)
	Ambient (Lab)	66 – 70 (˚F)
	Cold 39 - 43 (˚F)
Soil Moist.	Saturated	20 – 25%
	Wet	15 – 20%
	Moist	10 – 15%

To clarify abbreviations in future tables: “C” is cold, “R” is room temperature, “H” is hot, “M” is moist, “W” is wet, and “S” is saturated. These specific combinations were chosen during preliminary feasibility studies which indicated that these five combinations contain the widest range of corrosion rates in soil. EPRI collects corrosion rates using the linear polarization method with a Gamry Instruments® Potentiostat.

Data & Results

Corrosion Rates in Soil

Table 2 displays EPRI-developed categorization of severity of steel, zinc, and copper in soil exposure based on the soil library of approximately 800 laboratory-measured corrosion rates. These “R” designations simplify subgrade corrosion severity into categories to assist mitigation and maintenance schedules. These designations are also used populate EPRI’s subgrade corrosion maps, which will be briefly discussed later in this report. In Table 2, all values are in mils of loss per year (mpy). “R1” is minimally corrosive subgrade, increasing in severity to “R4”, which is extremely corrosive.

Table 2: Corrosion Rate Ranges for Metals in Soil.

	R1	R2	R3	R4
Steel	0.0–0.5	0.5–1.0	1.0–3.0	>3.0
Copper	0.0–0.18	0.18–0.38	0.38–1.15	>1.15
Zinc	0.0–0.35	0.35–0.85	0.85–2.83	>2.83

Laboratory analysis revealed that the native soils around the Substation are widely variable in terms of corrosion to steel, zinc, and copper (Figure 3, Figure 4, and Figure 5). The ranges are as follows:

• Copper - .07 to 1.6 mils per year
• Steel - .03 to 5.9 mils per year
• Zinc - .05 to 4.8 mils per year

It is important to recall the moisture and temperature ranges when understanding the performance of assets in these soils (Table 1). As can be seen in the performance of all metals, moisture content was the strongest dictator of corrosion severity. For example, note the small discrepancies between all measured “CM”/”HM” or “CS”/”HS”, versus the large discrepancies between “CM”/”CS” or “HM”/”HS”.

Soil Chemistry

In analytical chemistry for corrosion, the primary metrics of interest are chlorides, nitrates, and pH due to their known influence on corrosion rates. These data, as collected by the EPRI soil lab, are summarized in Table 3. As mentioned earlier, a comprehensive list of the evaluated organic and inorganic chemistry is summarized and described in Appendix A.

Table 3: Primary Corrosion-Driving Chemistry.

Soil ID	Chlorides (ppm)	Nitrates (ppm)	pH
FcC 1	5	4.91	5
FcC 2	9.62	4.57	5.82
FcC 3	5.64	3.67	5.17
NwB 1	3.86	4.57	5.69
NwB 2	3.49	4.68	4.92
NwB 3	3.9	19.1	3.82
PwB 1	4.94	3.92	5.22
PwB 2	0.79	1.11	6.33
PwB 3	0.625	1.04	7.46

These primary metrics provide an understanding of the soil conductivity and the acidity levels that may affect amphoteric metals such as lead, zinc, and copper. General corrosion (Figure 6L) on steel occurs more frequently in acidic soils (pH <7), while localized or pitting corrosion (Figure 6R) occurs more in neutral soils (pH ~7) due to pitting corrosion have greater penetration rates. In soils with high concentrations of chlorides or nitrates , corrosion levels are elevated due to the breakdown of the passivation layers on the various metals.

Figure 6: Examples of general corrosion (L) and pitting corrosion (R).

Soil Texture

To understand the significance of texture, soil ternary diagrams, or texture triangles, can visualize the relationship between particle distribution (percent sand, silt, and clay) and corrosion rate of copper, zinc, and steel. EPRI has developed three soil triangles to understand soil corrosivity and PSA based on the entire soil library database to establish general trends (Figure 7, Figure 8, and Figure 9).

Soil texture is a governing factor in time of wetness, and moisture is required to maintain a corrosion cell. Hence, well-drained soils are typically are less corrosive due to freely moving moisture and precipitation cleansing contaminants such as chlorides, nitrates, and sulfates from the soil that may accelerate corrosion. Clays hold moisture, but are considered anaerobic, which may retard the oxidation or corrosion process. Silts and loamy soils hold some moisture and are more porous and permeable in nature, therefore supporting higher corrosion rates. The soils collected at the substation were primarily categorized as loam or silt loam. Both categories are found near the bottom of the soil triangles and are corrosive in nature due to the ability to hold moisture.

Table 4: Substation Soil PSA.

Sample	% clay (<4um)	% sand (>62um)	% silt
FcC1	13.7	28.5	57.8
FcC2	12.5	32.1	55.4
FcC3	17	18.3	64.7
NwB1	9.9	47.5	42.6
NwB2	10.5	68.1	21.4
NwB3	11	31	58
PwB1	16	34.3	49.7
PwB3	10.9	45.8	43.3

Color

Another important metric to quickly quantify a soil as actively corrosive or benign is using color. The color of a soil can provide an understanding of corrosivity based upon oxygen content or moisture-retention capability. Generally, soil that is black or dark brown indicates high organic matter content, therefore higher acidity, and increased corrosion. A greyer or bluer soil indicates higher water content – with water availability, corrosion can increase (unless conditions are anoxic). Meanwhile, reddish or bright colors indicate relative non-corrosivity, being that they have already oxidized and are not actively corroding. The soils from the Substation were evaluated using the Munsell Color System, a hue/value/chroma color space (Table 5). Generally, soils were brown and dark, indicating rich organics and high acidity. The grayish colors are demonstrative of extended times of wetness.

Table 5: Color to Support Corrosivity Measurements.

Wet Hue	Wet Value	Wet Chroma	Name
10	3	2	Brown
10	3	2	Brown
10	4	1	Dark Gray
10	3	2	Brown
10	3	2	Brown
2.5	5	2	Grayish Brown
10	4	2	Dark Grayish Brown
2.5	4	2	Dark Grayish Brown
10	3	2	Brown

Determining Drivers

In the data analytics stage, a large multivariate analysis is run to screen for any statistically significant relationships (coefficient of determination [r2] of about 0.5 or greater) in this large data set. This screening is completed to understand what known corrosion drivers could be accelerating corrosion locally, and for which metal.

Zinc

Zinc soil corrosion rates had the strongest correlations with texture-related data. In the following graphs, it can be seen that there is a positive correlation between soil weight/volume ratio (see Appendix A for description) and corrosion rates (Figure 10), as well as strong correlations with percent clay, silt, and sand (Figure 11). This all affirms the findings of the electrochemical methods earlier that moisture and time of wetness is a primary driver of corrosion in this region.

Steel

In the case of the Substation, steel corrosion rates seem most sensitive to acid content in soils – based upon the correlations with pH and exchangeable acidity (Ac) (Figure 12). This is further indicated by the strong relationship with humic (organic) matter content (HM) (Figure 13). It can be anticipated that any exposed steel at groundline exposed to high organics/high acidity with have a decreased corrosion rate.

Copper

Copper corrosion rates did not display a particularly strong correlation with any specific measured parameter. Determination coefficients ranged from 0.25 – 0.45 for the five primary drivers of corrosion: pH, moisture, temperature, potential, and conductivity. This is a prime example of the complexity of soil corrosion – there is no one correct answer for what is driving corrosion at any particular time, because it is a combination of many influencers that require multifaceted monitoring and mitigation.

Conclusions

Soils collected at the substation were found to be highly corrosive to steel, zinc, and copper during times of wetness. These soils have the ability to hold enough moisture and enough oxygen to induce an active corrosion cell, so that periods of corrosion activity is extended. It should also be noted that cold temperatures significantly diminish the corrosion rates (Figure 14), so seasonal periods of active corrosion are limited to late spring, summer, and early fall.

The ambient conditions of all represented soil series (Room Wet conditions) are fairly comparable within 0.4 mpy. The net effect is an average corrosion rate for each soil sample as the moisture and temperature fluctuates annually (Figure 15).

When EPRI has measured the corrosion rate for a soil series, “R” designations are assigned to that soil series and then, using spatial interpolation in GIS, are projected to all locations where that soil series can be found. Using the average corrosion rates for zinc (Figure 16), steel (Figure 17), and copper (Figure 18), it was possible to generate preliminary subgrade corrosion maps for the Substation and the surrounding area. The map may seem fairly low-resolution at this scale, but the overall vision is that even this small sampling area has implications of interpolation throughout the utility territory. Light Green is “R2”, Orange is “R3”, and Red is “R4” for all metals.

The subgrade corrosion map currently resides in ArcGIS Online and EPRI is in the process of creating access points for utilities to view their territories directly.

An additional implementation of soil corrosion evaluation in a laboratory is the ability to create localized subgrade corrosion models. Based on EPRI’s extensive laboratory analyses, it has been determined that five measurable variables dominate the soil corrosion rate – temperature, moisture, conductivity, native potential, and pH. With this model comes a soil corrosion linear regression algorithm, for which the values for the five parameters governing the kinetics of corrosion can be input and produce an anticipated soil corrosion rate of a metal for that soil or group of soils. The following series of graphs and equations are the algorithmic models for Manchester Substation, and can be used for prediction of subgrade corrosion rates.

Equation 1: Substation Subgrade Corrosion Algorithm for Zinc
CR= -6.177+(-0.005OCP)+(-0.003Temp)+(0.127VWC)+(1.016EC)+(0.101*pH)

Where:

• CR: Corrosion Rate (mpy) OCP: Open Circuit Potential (mV)
• Temp: Soil Temperature (°C)
• VWC: Soil Water Content (%)
• EC: Electrical Conductivity (mS/cm)
• pH: Soil pH (unitless)

Equation 2: Substation Subgrade Corrosion Algorithm for Steel CR=-4.270+(-0.007OCP)+(0.001Temp)+(0.098VWC)+(-3.304EC)+(0.298*pH)

Equation 3: Substation Subgrade Corrosion Algorithm for Copper CR=-0.249+(-1.160e^(-6)OCP)+(0.004Temp)+(0.040VWC)+(-0.496EC)+(-0.021*pH)

Recommendations

To summarize the recommendations for this soil corrosivity study, the “R” designation table is revisited once again (Table 6).

Table 6: Corrosion Rate Ranges for Metals in Soil

	R1	R2	R3	R4
Steel	0.0–0.5	0.5–1.0	1.0–3.0	>3.0
Copper	0.0–0.18	0.18–0.38	0.38–1.15	>1.15
Zinc	0.0–0.35	0.35–0.85	0.85–2.83	>2.83

This is essential because EPRI has assigned recommended corrective actions for each of these categories. For the lowest corrosion rates, “R1,” no action is necessary. For “R2,” it is time to monitor any assets with this level of corrosion activity and begin trending and modelling. In the case of significant corrosion “R3,” it would be necessary to review and repair coating systems and apply cathodic protection. On top of all of this, in severe “R4” cases, a utility would also need to make structural repairs to any damaged components.

Also of note is the existence of tool-less, diagnostic subgrade corrosion field tests that can be performed to understand texture and color, including the utilization of the Munsell Color System and the USDA’s hand-feel soil texture test. Relationships between texture, color, and corrosion rates of soils around the Substation have been established by this soil characterization study. Utilizing similar diagnostics to obtain general corrosion knowledge in the field will assist in data collection and maintenance scheduling optimization.

It is recommended that corrosion control methods are deployed at the Substation to reduce or arrest corrosion on assets in contact with soil. Cathodic protection has been demonstrated to mitigate corrosion within substations, and to eliminate areas with corrosion due to circulating or stray currents. An impressed current cathodic protection system may be designed and implemented once the soil resistivity model has been completed, and copper mass of the ground grid has been calculated.

Further, more diagnostics are required to understand any circulating and stray currents within the substation contributing to subgrade corrosivity.

Appendix A

Table 7: Complete Chemical Content Measured in Soil Samples

All of the following descriptions are provided by the North Carolina Department of Agriculture. Original Document

Humic Matter (HM%) Humic matter percent is a measurement of humic and fulvic acid components of soil organic matter, expressed on a volume basis. It represents the percentage of soil organic matter that is soluble in a dilute alkaline solution (NaOH). This portion represents the chemically active organic fraction used in determining lime rates. Humic matter percent may or may not correlate with percent organic matter.

Weight/Volume Ratio (W/V) Weight/volume ratio, expressed in g/cm3 , is used to classify the soil type of a sample. For example, a very sandy soil may have a W/V of 1.5 g/cm3 , whereas the W/V of an organic soil may be as low as 0.5 g/cm3 . Soils high in clay fall within these two extremes. W/V is generally inversely related to the cation exchange capacity (CEC) of the soil: that is, soils with a high W/V generally have a low CEC.

Cation Exchange Capacity (CEC) CEC is a relative measure of the nutrient-holding capacity of a soil. It is expressed in units of meq/100 cm3 and is determined by summation of extractable calcium, magnesium, potassium, and exchangeable acidity (Ac). The CEC of North Carolina soils ranges from low (< 2.0 meq/100 cm3 ) for sandy soils to as high as 25 meq/100 cm3 for clay and organic soils. A high CEC is desirable because nutrients are less subject to leaching and adequate quantities of nutrient reserves can be maintained. However, sandy soils, by nature, have low CEC, and little can be done to change it. The CEC will vary with changes in soil pH, organic matter, and clay content.

Base Saturation (BS%) Base saturation is expressed as a percentage of the CEC that is occupied by basic cations, principally calcium, magnesium, and potassium. Base saturation and pH are directly correlated: as pH increases, so does BS%. A higher BS% also corresponds to a lower level of soil acidity. Exchangeable Acidity (Ac) Exchangeable acidity represents that portion of the CEC that is occupied by hydrogen (H+ ) and aluminum (Al+++) and is expressed as meq/100 cm3 .

Current pH The pH is a measure of the active acidity or the hydrogen ion (H+) activity in the soil solution.

3 - Substation Ground Grid Corrosion Study 2

Introduction

Gaps were identified within IEEE Std 81 “Guide for Measuring Earth Resistivity, Ground Impedance, and Earth Surface Potentials of a Grounding System” for identifying and locating corrosion damage within substation ground grids. To satisfy these gaps, EPRI has been developing methods based on soil characterization and electrochemistry to augment the existing methodology.

Objective

Soil characterization provides an understanding of soil corrosivity and how asset materials, such as copper, zinc (galvanizing), and steel, perform throughout long-term environmental exposure. Soil-exposed assets include the foundations for bus work, tank bottoms, and the ground grid itself. Results of soil characterization may be used in a host of applications that include site planning, forensics, and optimizing maintenance operations.

Background

EPRI was contacted by a utility to understand the corrosivity of soils throughout their territory, primarily focused on substation assets. EPRI reviewed the 205 substation assets first filtering by sampling site integrity (i.e. access, urbanization of soils), then by reviewing the soil “series” at each location. These locations were selected based on the NRCS Database “SSURGO”, where the US Soil Survey’s maps are compiled, delineating soil series taxonomic groupings throughout the country. When a soil has similar characteristics as defined by the soil surveyors, it is mapped as a polygon and labeled as a “series”. Sample locations at three substations (S429, S444, S594) were selected to be representative of the three most-occurring native soil series found at nearly 20% of substation asset locations (Figure 1). Each substation was sampled for two native soil series and the backfill used at that location (Figure 2). A LCRA field crew collected soil samples at these locations.

The collected soils were packaged, labeled, and shipped to the Charlotte Corrosion and Soils Laboratory for analysis. EPRI’s soil characterization consists of chemical analysis, particle size analysis, and linear polarization resistance using electrochemistry. The following metrics were measured during analysis:

• Nitrates
• pH
• Calcium Chloride
• Humic Matter
• Cation Exchange Capacity
• Base Saturation Percent
• Exchangeable Acidity
• Calcium
• Magnesium
• Manganese
• Zinc
• Sodium
• % Clay
• % Sand
• % Silt
• Color (Hue, Value, Chroma)
• Corrosion Rates for zinc, steel, and copper as a function of moisture and temperature

Methodology

The following is a summary methodology of EPRI’s soil corrosion evaluation procedures. For an extensive overview, please reference Recommended Practice Guidelines for Corrosion Inspection and Assessment (2020)

Chemistry

pH testing is based on ASTM D4972-13: Standard Test Method for pH of Soils. The pH of a soil is a useful variable in determining the solubility of soil minerals and the mobility of ions in the soil.

Nitrate concentration measurements are based on Colorado Department of Transportation Laboratory Procedure (CP-L 2104): Determining the chloride ion content in water or water-soluble chloride ion content in soil and Texas Department of Transportation Procedure (Tex-620-J): Determining chloride and sulfate contents in soil. Nitrates are extracted from the soil into a deionized water solution, which is then filtered to remove any remaining soil particulates. The actual measurement of the nitrate ion concentrations is completed using Ion Selective Electrodes (ISE) for each of the ion types. The remaining chemical analyses are completed by the North Carolina Department of Agriculture due to the availability and expediency of testing. The full table of chemistry results from this testing, along with descriptions of calculated parameters, can be found in Appendix A.

Particle Size Analysis (PSA)

EPRI evaluates particle size using a coarse sieve stack and Beckman-Coulter laser diffraction for the fine fraction distribution (Figure 3).

Corrosion Rates

EPRI collects corrosion rate data using an electrochemical tabletop testing method called MT5 to highlight a range of corrosive environmental conditions. The procedure was created to understand the impact of soil moisture and temperature on the corrosion rate of copper, steel, and zinc. In this procedure, a soil is subjected to varying moisture and temperature plateaus, then corrosion rates are measured for different metals. Electrical conductivity, pH, and native potential are also measured. Five combinations of temperature and moisture are measured in the MT5 procedure from Table 1: Cold Moist, Hot Moist, Room Wet, Cold Saturated, and Hot Saturated.

Table 1: MT5 Parameters.

Level	Range
Soil Temp.	Hot	104 - 111 (˚F)
	Ambient (Lab)	66 – 70 (˚F)
	Cold 39 - 43 (˚F)
Soil Moist.	Saturated	20 – 25%
	Wet	15 – 20%
	Moist	10 – 15%

To clarify abbreviations in future tables: “C” is cold, “R” is room temperature, “H” is hot, “M” is moist, “W” is wet, and “S” is saturated. These specific combinations were chosen during preliminary feasibility studies which indicated that these five combinations contain the widest range of corrosion rates in soil. EPRI collects corrosion rates using the linear polarization method with a Gamry Instruments® Potentiostat.

Data & Results

Corrosion Rates in Soil

Table 2 displays EPRI-developed categorization of severity of steel, zinc, and copper in soil exposure based on the soil library of approximately 670 laboratory-measured corrosion rates. These “R” designations simplify subgrade corrosion severity into categories to assist mitigation and maintenance schedules. These designations are also used populate EPRI’s subgrade corrosion maps, which will be briefly discussed later in this report. In Table 2, all values are in mils of loss per year (mpy). “R1” is minimally corrosive subgrade, increasing in severity to “R4”, which is extremely corrosive.

Table 2: Corrosion Rate Ranges for Metals in Soil.

	R1	R2	R3	R4
Steel	0.0–0.5	0.5–1.0	1.0–3.0	>3.0
Copper	0.0–0.18	0.18–0.38	0.38–1.15	>1.15
Zinc	0.0–0.35	0.35–0.85	0.85–2.83	>2.83

Laboratory analysis revealed that the native soils around the Substation are widely variable in terms of corrosion to steel, zinc, and copper (Figure 3, Figure 4, and Figure 5). The ranges are as follows:

• Copper - .07 to 1.6 mils per year
• Steel - .03 to 5.9 mils per year
• Zinc - .05 to 4.8 mils per year

It is important to recall the moisture and temperature ranges when understanding the performance of assets in these soils (Table 1). As can be seen in the performance of all metals, both temperature and moisture were strong corrosion drivers. For example, note the small discrepancies in zinc corrosion between measured “CM”/”HM” or “CS”/”HS”, versus the large discrepancies between “CM”/”CS” or “HM”/”HS”.

Soil Chemistry

In analytical chemistry for corrosion, some metrics of interest are nitrates and pH due to their known influence on corrosion rates. These data, as collected by the EPRI soil lab, are summarized in Table 3. As mentioned earlier, a comprehensive list of the evaluated organic and inorganic chemistry is summarized and described in Appendix A.

Table 3: Primary Corrosion-Driving Chemistry.

Soil ID	Nitrates (mg/kg)	pH
S429 1	4	8
S429 2	5	8
S429 3	5	8
S444 1	18	8
S444 2	6	8
S444 3	7	8
S594 1	4	8
S594 2	6	8
S594 3	11	8

These primary metrics provide an understanding of the soil conductivity and the acidity levels that may affect amphoteric metals such as lead, zinc, and copper. General corrosion (Figure 7L) on steel occurs more frequently in acidic soils (pH <7), while localized or pitting corrosion (Figure 7R) occurs more in neutral soils (pH ~7) due to pitting corrosion have greater penetration rates. In soils with high concentrations of nitrates , corrosion levels are elevated due to the breakdown of the passivation layers on the various metals.

Figure 7: Examples of general corrosion (L) and pitting corrosion (R).

Soil Texture

To understand the significance of texture, soil ternary diagrams, or texture triangles, can visualize the relationship between particle distribution (percent sand, silt, and clay) and corrosion rate of copper, zinc, and steel. EPRI has developed three soil triangles to understand soil corrosivity and PSA based on the entire soil library database to establish general trends (Figure 8, Figure 9, and Figure 10).

Soil texture is a governing factor in time of wetness, and moisture is required to maintain a corrosion cell. Hence, well-drained soils are typically are less corrosive due to freely moving moisture and precipitation cleansing contaminants such as chlorides, nitrates, and sulfates from the soil that may accelerate corrosion. Clays hold moisture, but are considered anaerobic, which may retard the oxidation or corrosion process. Silts and loamy soils hold some moisture and are more porous and permeable in nature, therefore supporting higher corrosion rates.

The soils collected at the LCRA substations were primarily categorized as loam or silt loam. Both categories are found near the bottom of the soil triangles and are corrosive in nature due to the ability to hold moisture and have some aeration – both elements essential to a corrosion cell.

Table 4: LCRA Substation Soil PSA.

Sample	% clay (<4um)	% sand (>62um)	% silt
S429 1	49.1	7.39	43.51
S429 2	22.1	27	50.9
S429 3	20.3	42.6	37.1
S444 1	19.4	45.8	34.8
S444 2	40.6	12.7	46.7
S444 3	32.6	14	53.4
S594 1	35.4	8.2	56.4
S594 2	18	47.7	34.3
S594 3	19.4	26.2	54.4

Color

Another important metric to quickly quantify a soil as actively corrosive or benign is using color. The color of a soil can provide an understanding of corrosivity based upon oxygen content or moisture-retention capability. Generally, soil that is black or dark brown indicates high organic matter content, therefore higher acidity, and increased corrosion. A greyer or bluer soil indicates higher water content – with water availability, corrosion can increase (unless conditions are anoxic). Meanwhile, reddish or bright colors indicate relative non-corrosivity, being that they have already oxidized and are not actively corroding. The soils were evaluated using the Munsell Color System, a hue/value/chroma color space (Table 5). Generally, soils were brown and dark, indicating rich organics and high acidity. The grayish colors are demonstrative of extended times of wetness. The bright yellows and reds colors are indicative of completed oxidation, therefore these soils show have low corrosion rates.

Table 5: Color to Support Corrosivity Measurements.

Wet Hue	Wet Value	Wet Chroma	Name
10YR	7	6	Yellow
10YR	2	1	Black
10YR	4	2	Dark Grayish Brown
10YR	8	3	Very Pale Brown
10YR	3	1	Very Dark Gray
10YR	2	1	Black
7.5YR	6	6	Reddish Yellow
7.5YR	5	3	Brown
10YR	2	2	Very Dark Brown

Determining Drivers

In the data analytics stage, a large multivariate analysis is run to screen for any statistically significant relationships (coefficient of determination [r2] of about 0.5 or greater) in this large data set. This screening is completed to understand what known corrosion drivers could be accelerating corrosion locally, and for which metal. For improvement of accuracy, backfill samples were excluded in multivariate analysis. Zinc Zinc corrosion rates did not display a particularly strong correlation with any specific measured parameter. Determination coefficients ranged from 0.22 – 0.53 for the five primary drivers of corrosion: pH (0.26), moisture (0.33), temperature (0.47), potential (0.53), and conductivity (0.22). This is a prime example of the complexity of soil corrosion – there is no one correct answer for what is driving corrosion at any particular time, because it is a combination of many influencers that require multifaceted monitoring and mitigation. Zinc soil corrosion rates had the strongest correlations with the external factors soil temperature and moisture variables (Figure 11), which affirms the results of the electrochemical testing data (Figure 6).

Steel

Similar to zinc corrosion, determination coefficients were fairly evenly spread throughout all measured variables. For the five primary drivers of corrosion, determination coefficients were as follows: potential (0.76), temperature (0.33), moisture (0.31), conductivity (0.24), and pH (0.22). Interestingly, steel corrosion rates displayed a positive correlation with sodium (Na) concentration rates in the soil (Figure 12). This may be attributed to the semi-aridity of the Texas climate – salts accumulate as evaporites on/near the surface of the soils, then are activated during periods of precipitation, making them highly corrosive for short durations.

Copper

Like the other metals, determination coefficients were fairly evenly spread throughout all measured variables. For the five primary drivers of corrosion, determination coefficients were as follows: potential (0.77), temperature (0.27), moisture (0.42), conductivity (0.41), and pH (0.43). A fairly notable relationship with sodium concentration was also observed (0.38). Copper corrosion’s most statistically significant driver in this region appears to be pH (Figure 13).

Conclusions

Soils were found to be increasingly corrosive to steel, zinc, and copper during times of wetness. These soils have the ability to hold enough moisture and enough oxygen to induce an active corrosion cell, so that periods of corrosion activity is extended. It should also be noted that cold temperatures significantly diminish the corrosion rates (Figure 14), so seasonal periods of active corrosion are limited to late spring, summer, and early fall.

The ambient conditions of all represented soil series (Room Wet conditions) are fairly comparable with some outlier exceptions. The net effect is an average corrosion rate for each soil sample as the moisture and temperature fluctuates annually (Figure 15).

When EPRI has measured the corrosion rate for a soil series, “R” designations are assigned to that soil series and then, using spatial interpolation in GIS, are projected to all locations where that soil series can be found. Using the average corrosion rates for zinc (Figure 16), steel (Figure 17), and copper (Figure 18), it was possible to generate preliminary subgrade corrosion maps for much of the surrounding area. The map may seem fairly low-resolution at this scale, but the overall vision is that even this small sampling area has implications of interpolation throughout the utility territory. Light Green is “R2”, Orange is “R3”, and Red is “R4” for all metals.

The subgrade corrosion map currently resides in ArcGIS Online and will be available online summer of 2022. An additional implementation of soil corrosion evaluation in a laboratory is the ability to create localized subgrade corrosion models. Based on EPRI’s extensive laboratory analyses, it has been determined that five measurable variables dominate the soil corrosion rate – temperature, moisture, conductivity, native potential, and pH. With this model comes a soil corrosion linear regression algorithm, for which the values for the five parameters governing the kinetics of corrosion can be input and produce an anticipated soil corrosion rate of a metal for that soil or group of soils. The following series of graphs and equations are the algorithmic models and may be used for prediction of subgrade corrosion rates within the surrounding territory. Backfill samples were excluded from modelling.

Equation 1: Substation Subgrade Corrosion Algorithm for Zinc
Zinc CR=-19.403+(-0.015OCP)+(0.039Temp)+(0.023VWC)+(-0.002EC)+(0.845*pH)

Where:

• CR: Corrosion Rate (mpy) OCP: Open Circuit Potential (mV)
• Temp: Soil Temperature (°C)
• VWC: Soil Water Content (%)
• EC: Electrical Conductivity (mS/cm)
• pH: Soil pH (unitless)

Equation 2: Subgrade Corrosion Algorithm for Steel. Steel CR=-2.843+(-0.003OCP)+(0.106Temp)+(0.233VWC)+(-0.008EC)+(-0.356*pH)

Equation 3: Subgrade Corrosion Algorithm for Copper. Copper CR=-2.663+(-0.007OCP)+(0.037Temp)+(-0.036VWC)+(-0.001EC)+(0.494*pH)

Recommendations

To summarize the recommendations for this soil corrosivity study, the “R” designation table is revisited once again (Table 6).

Table 6: Corrosion Rate Ranges for Metals in Soil

	R1	R2	R3	R4
Steel	0.0–0.5	0.5–1.0	1.0–3.0	>3.0
Copper	0.0–0.18	0.18–0.38	0.38–1.15	>1.15
Zinc	0.0–0.35	0.35–0.85	0.85–2.83	>2.83

This revisit is essential because EPRI has assigned recommended corrective actions for each of these categories. For the lowest corrosion rates, “R1,” no action is necessary. For “R2,” it is time to monitor any assets with this level of corrosion activity and begin trending and modelling. In the case of significant corrosion “R3,” it would be necessary to review and repair coating systems and apply cathodic protection. On top of all of this, in severe “R4” cases, a utility would also need to make structural repairs to any damaged components.

Also of note is the existence of tool-less, diagnostic subgrade corrosion field tests that can be performed to understand texture and color, including the utilization of the Munsell Color System and the USDA’s hand-feel soil texture test. Relationships between texture, color, and corrosion rates of soils around the substations and surrounding utility territory have been established by this soil characterization study. Utilizing similar diagnostics to obtain general corrosion knowledge in the field will assist in data collection and maintenance scheduling optimization.

It is recommended that corrosion control methods are deployed at the substations to reduce or arrest corrosion on assets in contact with soil. Cathodic protection has been demonstrated to mitigate corrosion within substations, and to eliminate areas with corrosion due to circulating or stray currents. An impressed current cathodic protection system may be designed and implemented once the soil resistivity model has been completed, and copper mass of the ground grid has been calculated.

Further, more diagnostics are required to understand any circulating currents within the substation contributing to subgrade corrosivity.

Appendix A

Table 7: Complete Chemical Analyses.

All of the following descriptions are provided by the North Carolina Department of Agriculture. Original Document

Humic Matter (HM%) Humic matter percent is a measurement of humic and fulvic acid components of soil organic matter, expressed on a volume basis. It represents the percentage of soil organic matter that is soluble in a dilute alkaline solution (NaOH). This portion represents the chemically active organic fraction used in determining lime rates. Humic matter percent may or may not correlate with percent organic matter.

Weight/Volume Ratio (W/V) Weight/volume ratio, expressed in g/cm3 , is used to classify the soil type of a sample. For example, a very sandy soil may have a W/V of 1.5 g/cm3 , whereas the W/V of an organic soil may be as low as 0.5 g/cm3 . Soils high in clay fall within these two extremes. W/V is generally inversely related to the cation exchange capacity (CEC) of the soil: that is, soils with a high W/V generally have a low CEC.

Cation Exchange Capacity (CEC) CEC is a relative measure of the nutrient-holding capacity of a soil. It is expressed in units of meq/100 cm3 and is determined by summation of extractable calcium, magnesium, potassium, and exchangeable acidity (Ac). The CEC of North Carolina soils ranges from low (< 2.0 meq/100 cm3 ) for sandy soils to as high as 25 meq/100 cm3 for clay and organic soils. A high CEC is desirable because nutrients are less subject to leaching and adequate quantities of nutrient reserves can be maintained. However, sandy soils, by nature, have low CEC, and little can be done to change it. The CEC will vary with changes in soil pH, organic matter, and clay content.

Base Saturation (BS%) Base saturation is expressed as a percentage of the CEC that is occupied by basic cations, principally calcium, magnesium, and potassium. Base saturation and pH are directly correlated: as pH increases, so does BS%. A higher BS% also corresponds to a lower level of soil acidity.

Exchangeable Acidity (Ac) Exchangeable acidity represents that portion of the CEC that is occupied by hydrogen (H+ ) and aluminum (Al+++) and is expressed as meq/100 cm3 .

Current pH The pH is a measure of the active acidity or the hydrogen ion (H+) activity in the soil solution.

4 - Corrosion Monitoring System

Originally designed as a research tool but has now found a home in monitoring “at risk” structures and anomalies that remain unexplained. The Corrosion Monitoring System (CMS) is a self-contained system that has power management, power supply, data storage, cellular connectivity and may be expanded to collect both atmospheric and subgrade data (see Figure 1).

The objectives of deploying the CMS is threefold, the first is to understand why assets degrade but also discriminate between corrosion types, the second is to locate suspect areas containing “at risk” assets and the third is to provide environmental data to support degradation models. Models consist of the following data and may be easily charted to trend both corrosion rates but also changes in subgrade and atmospheric weather (see Figure 2). The corrosion rates may be gathered for steel, zinc and copper for subgrade models or steel, zinc, aluminum, and copper in atmospheric models (see Figure 3).

EPRI has developed extensive atmospheric and soil corrosivity maps that may be augmented using data from the CMS data sets. Many departments within the utility may then use this new information for more accurate and informed life cycle decisions (see Figure 4).

The CMS data sampling rate may be set for a very high- or low-resolution dependent upon the application and how often the anomaly or environment changes. Figure 5 illustrates a Corrosion Monitoring System collecting subgrade data to determine if a cathodic protection system is performing properly and meeting the acceptance criteria of -850 mV or 100 mV shift.

Reviewing the chart, we may see that all of the bonded coupons are polarized more negative than -850 mV which is the acceptance criteria recommended by AMPP (formerly known as NACE). This polarization level allows the stake owners to determine that their asset is well protected and may provide many years of service.