Parameter Study of Financial Analysis for Implementing Solar Photovoltaics Structural Snow Fences

Abstract

Structural snow fences are known as a cost-effective way to enhance road safety on highways, which, however, are only used during winter, making them “useless” during summer. To increase their cost-effectiveness, Photovoltaics Snow Fences (PVSF) were developed by integrating PV panels with conventional structural snow fences. As part of the feasibility study supported by the Minnesota Department of Transportation (MnDOT), a financial analysis was performed involving many parameters, such as the capital and operating costs of the PVSF system, installation orientation of the panels, discount rates, energy selling prices, availability of incentives, ownership of the PV system, etc. The effects of these parameters on the analysis results were evaluated, where critical (most sensitive) parameters were first identified, and then their quantitative effects on the analysis results were evaluated in terms of Net Present Value (NPV) and Internal Rate of Return (IRR). The results indicate that the real discount rate is the most sensitive parameter in determining the cost-effectiveness of a PVSF project in Minnesota by looking at its NPV, when the benefits, such as Federal Tax Credits, Renewable Energy Certificates, and those associated with the use of snow fences, are considered in the financial analysis. The cost of the PVSF system is the most sensitive parameter for IRR, depending on the ownership of the PV system (by MnDOT or via a Power Purchase Agreement).

1. Introduction

Structural snow fences are typically used to eliminate blowing and drifting snow on highways. Specifically, the benefits of using snow fences along highways in cold climates include preventing snow drifting and blowing ice, as well as avoiding crashes and reducing travel time and carbon emissions. Recent cost-benefit studies on structural snow fences indicate that a 33% reduction in snow removal costs and a 70% reduction in crash rates can be achieved for a structural snow fence built along highways in Wyoming [1,2], and a project involving 2.25-km structural snow fences to be installed in Illinois has the potential to return the full cost in 5 years based on a cost-benefit analysis [3].

Nevertheless, structural snow fences are only used during winter, making them “useless” during summer. Integrating solar photovoltaics (PV) with structural snow fences adds more value to them and thus makes them useful not only in winter to reduce the amount of snow falling on highways but also in summer to generate electrical power using solar energy, which contributes to environmental sustainability. PV systems are used to convert solar energy, the most abundant renewable energy source, directly into electricity. Solar PV made up less than 0.1% of the world’s energy supply in 2005 [4] and less than 0.26% of all renewable energy sources [5], and the International Energy Agency estimates that its share of global generation will be only 2% in 2035 [6]. One of the challenges for implementing PV panels on a large scale is the limited available area. As a result, it may be feasible to combine PV with alternative applications that do not require additional space. Previous research demonstrated the benefit of combining PV panels on rooftops, building integrated PV, agrovoltaic systems, PV sited in low-land-quality areas, and PV noise barriers. [7]. This study takes a different approach by combining PV panels with snow fences for this purpose.

The integration of PV panels with conventional structural snow fences is called PV Snow Fences (PVSF) and is accomplished by replacing the snow fence’s rails between poles with PV panels, as shown in Figure 1. The dimension of the PV panels is expected to be the same as that of the rails to ensure that the original function of the structural snow fences will not be negatively affected. Figure 1 shows a 2.4-feet-tall 50% porosity PVSF as an example.

The idea of using PV panels near highways is not new. Switzerland was the first country to integrate PV panels with noise barriers, known as a PV Noise Barrier (PVNB), around 30 years ago in 1989 [8,9]. The US, however, lags behind the European countries in developing PVNBs along highways, and the first solar highway project in the US was developed by the Oregon Department of Transportation (ODOT) in 2008 [10], where the PV panels were not integrated with other facilities, such as noise barriers, but installed on the ground of the highway right-of-way, similar to a small-scale solar farm. Until recently, several pilot PVNB projects were considered by other state DOTs [9] to further advance the development of PV-integrated highway facilities in the US.

The development of PVSF provides the US an opportunity of a “corner overtaking” in developing highway PV projects. The concept of PVSF is innovative and its feasibility analysis was carried out via a recent research study funded by the Minnesota Department of Transportation (MnDOT). This paper is intended to summarize the findings of the study by focusing on the parameter study of a financial analysis performed to evaluate the cost-effectiveness of a PVSF project. All other aspects of the MnDOT-funded research project, e.g., the structural design of the PVSF or its electrical system development, would be reported separately in other publications. The findings reported in this paper are expected to be useful for the further development and implementation of PVSF projects, especially for their financial decisions, in the states of the US that are located in cold climates and potentially use or will use structural snow fences.

According to a literature review, there were quite a few studies that evaluated the costs and benefits of PV systems and snow fences, but none of them conducted a financial analysis considering both PV systems and snow fences together.

Tabler and Furnish in 1982 conducted the first analysis of the costs and benefits of installing a structural snow fence on a 100-km stretch of Interstate 80 in Wyoming, between Laramie and Walcott, considering snow removal expenditures, crash frequency, and road closure criteria. The study found that structural snow fences can reduce snow drifting by 33%, lower snow removal costs, and also cut crash rates by 70% when there is blowing snow. The installation cost of the snow fence was predicted to be amortized after 10 years of construction due to the decreased winter maintenance costs and savings on property damage [1].

In 2021, Baral et al. used a benefit-cost analysis (BCA) case study to examine and compare the benefit-cost ratios of living snow fences (LSFs), structural snow fences (SSFs), and standing corn rows (SCRs) on a 2.1 km blowing snow segment in McLean County (District 5) near Hudson, Illinois over a 15-year period in accordance with Federal Highway Administration (FHWA) guidelines. The BCA result demonstrated the cost-effectiveness (benefit-cost ratio > 1) of all three types. However, the ratios were lower than those in earlier studies because they considered the cost of leasing land for the snow fences, inconvenience costs, and a higher fee for snow removal [3].

Poullikkas (2009) carried out a feasibility analysis to determine whether the construction of PV parks in Cyprus is an economically viable option in the absence of an acceptable feed-in tariff or other measures. A parametric cost-benefit analysis was conducted by adjusting variables such as PV park orientation, PV park capital investment, discount rate, CO₂ emission trading system, and monthly fuel price to determine the least cost-feasible alternative for the installation of the PV park. The cost or benefit of the electricity unit before taxes, after-tax cash flow, net present value (NPV), internal rate of return (IRR), and payback period, were all computed for each case. The findings showed that, in the absence of feed-in tariffs, capital investment in the PV park is a crucial factor in determining the project’s profitability [11].

Gibson and Harder (2010) investigated the viability of constructing a 10 MW solar power plant in Abu Dhabi using RET Screen modeling software to predict energy output, financial feasibility, and greenhouse-gas (GHG) emission reductions. The findings revealed that the main reasons for not using PV systems in Abu Dhabi were their high initial costs and low predicted price for electricity generated, and they demonstrated that government incentives are required to make these projects feasible and accelerate the construction of PV power plants in Abu Dhabi, and a feed-in tariff rate of USD 0.16/kWh is suggested [12].

In this paper, we performed a financial analysis considering both PV systems and snow fences together, i.e., the PVSF, which involves many parameters (will be discussed in the next section). The effects of these parameters on the analysis results are the key to ensure the validity of the financial analysis model by properly handling the uncertainties associated with these parameters. A parameter study is, therefore, necessary to quantify the effects of these parameters and to do so, critical (most sensitive) parameters were first identified, which were then evaluated by quantifying their effects on the analysis results, in terms of Net Present Value (NPV) and Internal Rate of Return (IRR). This allows us to better understand the analysis model for further improvement and optimization. Additionally, this study can be used by decision makers to decide whether additional information is needed to narrow the range of estimates of some critical parameters for a more formal risk and uncertainty analysis [13]. Specifically, this paper consists of five sections, which are the introduction, financial analysis model, critical parameters identification, results, and conclusions, including future work.

2. Financial Analysis Model

The financial analysis augmented by consideration of social benefits consists of two separate sub-models for snow fences and the PV systems to be installed on them, respectively. The sub-model for the PV system considers all the direct and indirect capital costs and the Operation and Maintenance (O&M) costs, as well as the monetary benefits associated with the use of the PV system, such as the economic benefit when the generated energy is used to generate revenue by selling the generated power to utility companies and the environmental benefits when the reduced use of fossil fuels to generate the same amount of electrical energy and/or the associated reduction of GHG emissions are considered. The sub-model for snow fences considers their installation and material costs, land costs, O&M costs, and recycling costs, along with the benefits of using snow fences.

Table 1. Cost and benefit of the PVSF system.

Costs	Benefits
Total Costs = Direct Capital Costs + Indirect Capital Costs + Annual Costs + Future Costs	Total Benefits = Benefits Due to the PV System + Benefits Due to Snow Fences
Direct Capital Costs PV module Inverter Snow fence material Balance-of System (BOS) equipment Direct installation labor—PV system Direct installation labor—snow fences Grid interconnection and transmission Supply chain	Benefits due to the PV system Incentives, Federal Solar Investment Tax Credit (ITC), other incentives, etc. Renewable Energy Certificates (RECs) Revenues Generated by selling solar power to utility companies Electricity cost savings due to the lower electricity price via a Power Purchase Agreement (PPA) GHG emission reduction Salvage value
Indirect Capital Costs Permitting and environmental studies Customer acquisition and system design Other overheads Sales and income taxes
Annual Costs Inverter replacement Insurance Land cost for snow fences O&M costs for snow fences and PV system	Benefits due to snow fences Drifting savings Blow ice savings Avoided crashes Avoided travel time Avoided carbon emissions Salvage value
Future Costs Recycling cost

summarizes the costs and benefits of the PVSF system, and

Table 2. Cost information for the PVSF system.

PV System Cost Information		Comments/Notes
Direct Capital Costs
Module Price [USD/W]	USD 0.85	Customized PV [14]
Inverter Price [USD/W]	USD 0.15	[15]
Balance-of-System (BOS) Equipment [USD/W]	USD 0.35	[15]
Direct Installation Labor [USD/W]	USD 0.20	[15]
Grid Interconnection and Transmission [USD/W]	USD 0.05	[15]
Supply Chain Costs [% of the total material cost]	10%	[16]
Total Direct Capital Costs [USD/W]	USD 1.74	-
Indirect Capital Costs
Permitting and Environmental Studies [USD/W]	USD 0.05	[15]
Customer Acquisition and System Design [USD/W]	USD 0.05	[15]
Other Overheads [USD/W]	USD 0.20	[15]
Sales Taxes [%]	6.875%	[17]
Federal/State Income Tax Rate [%]	0.00%	MnDOT does not pay income taxes
Total Indirect Capital Costs [USD/W]	USD 0.38	-
Annual Costs
Inverter Lifetime [Years]	13	[15]
Inverter Replacement [USD/W]	USD 0.09	USD 0.09/W for small-scale PVSF systems [16]
Insurance Cost by Capacity [USD/kW-yr]	USD 5.00	USD 5.00/kW for small-scale PVSF systems [18]
O&M Annual Cost by Capacity [USD/kW-yr]	USD 10.00	USD 10.00/kW for small-scale PVSF systems [15]
Total O&M Costs [USD/W]	USD 0.38	USD 0.38/W for small-scale PVSF systems
Future Costs
Recycling Cost [USD/W]	USD 0.17	[19]
Structural Snow-Fence Cost Information	Comments/Notes
Install & Material Costs [USD/m]	USD 236.55	For 2.4-m-tall 50% porosity snow fences (Provided by MnDOT)
Land Cost [USD/linear m/year]	USD 3.28	Rental cost (Provided by MnDOT)
O&M Cost [USD/km-Year]	USD 1875	Provided by MnDOT
Recycling Cost [USD/m]	USD 0.82	[20]

and

Table 3. Benefit information for the PVSF system.

PV System Benefit Information		Comments/Notes
Federal ITC or other incentives	10.00%	For a project after 2022 [21]
RECs [USD/REC or USD/MWh]	USD 0.65	[22]
Price to sell back to a utility company [USD/kWh]	USD 0.11	According to survey responses from local utility companies
PPA for 3rd-party ownership [USD/kWh]	USD 0.10	[23]
System Salvage Value [% of Capital Cost]	15%	[24]
Greenhouse gas (GHG) Emission Cost Savings
Per metric ton CO2	USD 42.00	[25,26]
Per metric ton CH4	USD 1200.00	[25,26]
Per metric ton N2O	USD 15,000.00	[25,26]
GHG Emission Ratio (Coal/Solar)	8.37	[27]
Structural Snow-Fence Benefit Information		Comments/Notes
Drifting Savings [USD/km-Year]	USD 21,553.77	Agency cost savings with drifting-snow events (MnDOT data)
Blow Ice Savings [USD/km-Year]	6379.43	Agency cost savings with blowing snow and ice events (MnDOT data)
Avoided Crashes [USD/km-Year]	18,523.75	Cost savings from fatal, injury, and property-damage crashes (MnDOT data)
Avoided Travel Time [USD/km-Year]	8016.83	Savings caused by travel-time reductions due to improved road conditions (MnDOT data)
Avoided Carbon Emissions [USD/km-Year]	USD 150.88	Cost savings from reduced carbon emissions from the agency’s equipment (MnDOT data)
Salvage Value [USD/m]	USD 0.30	For steel poles [28]

detail the cost and benefit information used in the model.

Cash flow models were established by using the data in Table 2 and Table 3, along with the financial parameters and other data described in

Table 4. Financial parameters for cash flow models.

Financial Parameter Information		Comments/Notes
PVSF System’s Lifetime [Year]	25	Same as the PV system’s lifetime [29]
Real Discount Rate [%/yr]	3%	According to Federal Highway Administration (FHWA) recommendation [3,30]
Current Electricity Utility Price Paid by MnDOT [USD/kWh]	USD 0.12	According to the utility costs provided by MnDOT

. Specifically, the cash flows for the PV panels and the structural snow fences are integrated mathematically by adding the corresponding costs and benefits incurred at the same time (same year) together. Since the rails of structural snow fences will be replaced with the customized PV panels, the cost of the rails that accounts for about 18% of the total snow fence cost is excluded, and instead, the cost of the customized PV panels (Figure 1) was included to reflect the mutual influence when integrating these two cash flows.

A Power Purchase Agreement (PPA) is a financial mechanism, where a solar project’s developer procures, builds, operates, and maintains a solar system, while an organization accrues the benefits of solar power without owning the system by buying the power generated from the developer at a negotiated rate [23]. Therefore, the established financial analysis model was separated into two categories depending on the ownership of the PV system to be built, i.e.,

• Category 1: the PV system is owned by the organization, like MnDOT in the project,
• Category 2: the PV system is owned by a 3rd party organization (a solar developer) via a PPA.

This allows us to look for alternative solutions considering the different financial mechanisms.

3. Critical Parameter Identification

Critical parameters in the financial analysis represent those that have relatively large impacts on the analysis results when changed within a reasonable range, compared to other parameters whose impacts could be ignored. To identify those critical parameters, a list of the parameters to be identified (

Table 5. Critical parameter candidates.

Parameter Index	Parameters	Default Parameter Value	Varying Range		Comments
			Max	Min
PVSF-1	Total capital costs of the PV system [USD/W]	USD 2.13	USD 3.20	USD 1.07	Changed by up to ±50%: Considering the existence of uncertainties and the possible variations of PV-system costs in the future with the further development of solar technology.
PVSF-2	Annual costs of the PV system [USD/W]	USD 0.38	USD 0.57	USD 0.19
PVSF-3	Total capital costs of snow fences [USD/m]	USD 236.55	USD 354.83	USD 118.28	Changed by up to ±50%: Considering the existence of uncertainties and the possible variations of snow fence costs in the future.
PVSF-4	Land costs of snow fences [USD/Linear meter/year]	USD 3.28	USD 4.92	USD 1.64
PVSF-5	O&M costs of snow fences [USD/km/year]	USD 1875.00	USD 2812.5	USD 937.50
PVSF-6	Federal ITC and other incentives	10%	0%	22%	Considering that the ITC or other incentives may vary year after year [31].
PVSF-7	RECs [USD/MWh]	USD 0.65	USD 0	USD 1.10	Considering that utility companies may purchase the RECs to meet RES compliance [22].
PVSF-8	Price to sell back to utility company [USD/kWh]	USD 0.11	USD 0.03	USD 0.11	Considering that the price may vary among utility companies [32].
PVSF-9	PPA pricing [USD/kWh]	USD 0.10	USD 0.15	USD 0.05	Changed by up to ±50%: Considering that the PPA rate may vary among solar developers.
PVSF-10	Snow fence benefit (drifting, blow ice, avoided crashes, travel time, and carbon emissions) [USD/km/year]	USD 54,625	USD 54,625	USD 27,313	Reduced by up to 50%: Considering that the severity of the blowing-snow problem area with regards to the frequency and severity of crashes, the highway’s traffic volume, and the number of resources spent on materials/labor/equipment to keep the highway open may vary from site to site.
PVSF-11	Real discount rate [%/yr]	3%	7%	0%	Considering its impact of up to 7% [3,30].
PVSF-12	Electricity utility price paid by MnDOT [USD/kWh]	USD 0.12	USD 0.14	USD 0.10	According to the utility costs provided by MnDOT for its rest areas and/or street lights.
PVSF-13	Orientation [deg]	180°	270°	90°	Corresponding to different orientations, i.e., East (90°), South (180°), and West (270°).

) was generated, which were extracted from Table 2, Table 3 and Table 4. Table 5 shows the parameters along with the max. and min. values of these parameters that can vary. This is to set boundaries for those parameters to ensure that they can vary within the boundaries and they are not too large or small to lose their economic definitions and meanings when used in the model. To estimate the electrical power generated by using the proposed PVSF system (Figure 1), parameters in

Table 6. PVSF power generation.

PVSF		Comments/Notes
Panel Capacity [Watt/panel]	100	Customized PV panel to fit the snow fences as shown in Figure 1
Panel Length [m]	3.66	Same size as the snow-fence rail of a 2.4-m-tall, 50% porosity fence
Panel Width [m]	0.15	Same size as the snow-fence rail of a 2.4-m-tall, 50% porosity fence
Number of Panels per 1 mile or 1.6 km	3520	Eight customized panels (0.15 × 3.66 m) per section, installed vertically on snow fences (Figure 1)
Degradation Rate [%]	0.8	[33]
System Losses [%]	14.08	[15]
Tilt [deg]	90	Vertical installation as shown in Figure 1
Inverter Efficiency [%]	96	[15]

were used for a 1.6-km (or 1-mile) PVSF system, including 3520 customized PV panels that have the size of 0.15 × 3.66 m.

The sensitivity of each parameter listed in Table 5 was then evaluated, whose results in terms of NPV and IRR are shown in

Table 7. Sensitivity results for Category 1.

PVSF Parameters	NPV
	Mean	SD	Range	Diff%	Rank
1	USD 964,427.38	USD 269,804.98	USD 662,346.26	51.48%	2
2	USD 959,659.54	USD 40,411.48	USD 91,328.48	9.06%	9
3	USD 967,341.83	USD 126,311.90	USD 312,164.16	27.88%	5
4	USD 972,439.40	USD 36,027.67	USD 91,941.42	9.11%	8
5	USD 969,555.78	USD 20,347.92	USD 52,239.44	5.28%	10
6	USD 963,704.96	USD 59,157.20	USD 149,657.20	14.41%	7
7	USD 962,688.00	USD 2156.93	USD 5649.78	0.59%	11
8	USD 737,713.80	USD 242,576.39	USD 613,675.12	46.90%	3
9	USD 963,704.96	0	0	0.00%	12
10	USD 796,733.38	USD 134,862.52	USD 339,313.08	35.21%	4
11	USD 1,031,271.16	USD 604,419.66	USD 1,466,507.08	81.67%	1
12	USD 963,704.96	0	0	0.00%	12
13	USD 889,923.17	USD 101,038.85	USD 186,358.22	19.34%	6
PVSF Parameters	IRR
	Mean	SD	Range	Diff%	Rank
1	11.49%	4.26%	10.41%	60.14%	1
2	10.55%	0.28%	0.63%	5.79%	8
3	10.79%	1.80%	4.43%	33.82%	2
4	10.64%	0.25%	0.64%	5.87%	7
5	10.62%	0.14%	0.36%	3.35%	9
6	10.62%	0.83%	2.09%	17.86%	5
7	10.58%	0.01%	0.04%	0.38%	10
8	8.93%	1.32%	3.33%	31.50%	3
9	10.58%	0.00%	0.00%	0.00%	11
10	9.40%	0.96%	2.42%	22.91%	4
11	10.58%	0.00%	0.00%	0.00%	11
12	10.62%	0.00%	0.00%	0.00%	11
13	9.94%	0.64%	1.34%	12.67%	6

for Category 1 (owned by MnDOT) or

Table 8. Sensitivity results for Category 2.

PVSF Parameters	NPV
	Mean	SD	Range	Diff%	Rank
1	USD 1,245,668.82	0	0	0.00%	9
2	USD 1,245,668.82	0	0	0.00%	9
3	USD 1,249,305.68	USD 126,311.90	USD 312,164.16	22.27%	4
4	USD 1,254,403.25	USD 36,027.66	USD 91,941.42	7.12%	6
5	USD 1,251,519.63	USD 20,347.92	USD 52,239.44	4.11%	7
6	USD 1,245,668.82	0	0	0.00%	9
7	USD 1,245,668.82	0	0	0.00%	9
8	USD 1,245,668.82	0	0	0.00%	9
9	USD 1,290,867.05	USD 212,899.99	USD 564,977.92	36.97%	2
10	USD 1,593,990.64	USD 134,862.52	USD 339,313.09	27.24%	3
11	USD 1,301,044.57	USD 490,130.52	USD 1,189,109.19	61.89%	1
12	USD 1,245,668.82	USD 89,330.85	USD 225,991.17	16.63%	5
13	USD 1,226,472.23	USD 26,288.34	USD 48,486.78	3.89%	8
PVSF Parameters	IRR
	Mean	SD	Range	Diff%	Rank
1	28.80%	0.00%	0.00%	0.00%	8
2	28.80%	0.00%	0.00%	0.00%	8
3	34.03%	15.81%	38.84%	67.21%	1
4	28.73%	0.71%	1.71%	5.77%	5
5	28.87%	0.36%	0.97%	3.31%	7
6	28.80%	0.00%	0.00%	0.00%	8
7	28.80%	0.00%	0.00%	0.00%	8
8	28.80%	0.00%	0.00%	0.00%	8
9	29.69%	4.19%	11.12%	32.36%	2
10	25.67%	2.53%	6.36%	22.07%	3
11	28.80%	0.00%	0.00%	0.00%	8
12	28.58%	1.83%	4.45%	14.35%	4
13	28.34%	0.45%	0.96%	3.33%	6

for Category 2 (via a PPA). Table 7 and Table 8 include the mean, Standard Deviation (SD), range, and the % difference determined by varying one parameter from its minimum to maximum values, as defined in Equation (1) [34].

$$Diff=\frac{S_{max}-S_{min}}{S_{max}}$$

(1)

where, $Diff$ is the % difference, $S_{max}$ is the maximum output value, and $S_{min}$ is the minimum output value, when a parameter listed in Table 5 varies between the max. and min. boundaries. The % difference was used as the sensitivity index to determine parameter sensitivity, and the rank shown in Table 7 and Table 8 was determined according to it.

The results in Table 7 and Table 8 were also plotted in Figure 2 and Figure 3, respectively, which are clearer to show which parameters have significant impacts on the NPV and/or IRR. The x-axis in Figure 2 and Figure 3 represent the change between the min. and max. values of the different parameters from PVSF-1 to 13. These sensitivity analysis results were then used to identify the most sensitive and insensitive parameters that affect project NPV or IRR under both Categories 1 and 2.

4. Results and Discussions

As shown in both Table 7 and Table 8 and Figure 2 and Figure 3, for Category 1, where the PV system is owned by the organization, such as MnDOT, the top 5 most sensitive parameters for NPV are PVSF—11, 1, 8, 10, and 3, and for IRR are PVSF—1, 3, 8, 10, and 6, as indicated in

Table 9. Critical parameters for Category 1.

Rank	NPV		IRR
	Parameter Index	Parameters	Parameter Index	Parameters
Top 1	PVSF-11	Real discount rate	PVSF-1	Total Capital Costs of the PV system
Top 2	PVSF-1	Total Capital Costs of the PV system	PVSF-3	Total Capital Costs of Snow Fences
Top 3	PVSF-8	Price to Sell Back to Utility Company	PVSF-8	Price to Sell Back to Utility Company
Top 4	PVSF-10	Snow Fence Benefit	PVSF-10	Snow Fence Benefit
Top 5	PVSF-3	Total Capital Costs of Snow Fences	PVSF-6	Federal ITC and Other Incentives

.

Table 10. Critical parameters for Category 2.

Rank	NPV		IRR
	Parameter Index	Parameters	Parameter Index	Parameters
Top 1	PVSF-11	Real discount rate	PVSF-3	Total Capital Costs of Snow Fences
Top 2	PVSF-9	PPA Pricing	PVSF-9	PPA Pricing
Top 3	PVSF-10	Snow Fence Benefit	PVSF-10	Snow Fence Benefit
Top 4	PVSF-3	Total Capital Costs of Snow Fences	PVSF-12	Electricity Utility Price Paid by MnDOT
Top 5	PVSF-12	Electricity Utility Price Paid by MnDOT	PVSF-4	Land Cost of Snow Fences

shows the top 5 parameters for Category 2, where the PV system is owned by a third-party organization via a PPA.

As shown in Table 9, i.e., for the case scenario of Category 1, PVSF 11—Real Discount Rate is listed as the most critical (sensitive) parameter to the financial analysis result of NPV, which thus plays a significant role in determining the cost-effectiveness of a PVSF project. A relatively low real-discount rate, e.g., caused by COVID 19 [35], would help a PVSF project to be more cost-effective compared with a project with a higher rate, as shown in Figure 2a, if the impact of the pandemic would last for years with such a low rate. Unsurprisingly, PVSF 1, 8, 10, and 3 are listed as the 2nd, 3rd, 4th, and 5th sensitive parameters, respectively, in Table 9 in terms of NPV. For PVSF 1 (Total Capital Costs of the PV System), it varies over time, and in fact it decreases year after year in the US. For example, the commercial rooftop PV cost benchmark in 2020 was USD 1.74/W_DC, and in 2021 it dropped to USD 1.56/W_DC [36]. It is believed that this trend with lower PV system costs year after year contributes to the cost-effectiveness of a PVSF project. For PVSF 8—Price to Sell Back to Utility Company, it depends not only on utility companies, but also may be subject to the net metering cap of, e.g., 40 kW, i.e., a lower energy selling price, such as USD 0.02~0.03/kWh, would be involved for a project with 40 kW or greater [37]. Snow Fence Benefit (PVSF 10) also varies considering that the severity of the blowing-snow problem area with regards to the frequency and severity of crashes, the highway’s traffic volume, and the number of resources spent on materials/labor/equipment to keep the highway open may vary from site to site. A reasonable estimation from historical data is a must to ensure the accuracy of the financial analysis result, not only for the MnDOT, but also for other state DOTs. Additionally, the lower costs of structural snow fences (PVSF 3) play a significant role, even though their costs do not vary as much as PV systems do.

The top 5 sensitive parameters for IRR are a bit different since PVSF 11 does not change when determining IRR. IRR is typically used to estimate the profitability of a PVSF project [38]. As shown in Table 9, PVSF 1, 3, 8, and 10 are still listed as the critical parameters to determine the profitability of a project. In addition, PVSF 6 (Federal ITC and Other Incentives) would influence the decision to invest in a PVSF project. The Investment Tax Credit (ITC) provides tax relief for a solar project, and under the ITC, the Internal Revenue Service (IRS) approves a tax credit that is equal to 26% of the project’s total cost for the year of 2022 and then reduced to 22% in 2023 [39]. Database of State Incentives for Renewables & Efficiency (DSIRE) is a good source to examine when searching for policies and incentives by state [40]. A PPA allows MnDOT to benefit indirectly from tax incentives through lower electricity prices by using tax equity [23], considering that MnDOT generally does not pay taxes. This would thus make Category 2 (Table 10) more attractive for DOTs.

As shown in Table 10, when the PV system to be installed on snow fences is owned by a 3rd-party organization (a solar developer) through a PPA, PVSF 11, 10, and 3 are still listed as the top critical parameters/factors, but instead of PVSF 1 and 8 (listed in Table 9), PVSF 9 and 12 are more critical in decision making in terms of the investment of a PVSF project. The PPA pricing (PVSF 9) is the electricity price when MnDOT purchases the electrical power generated from the PVSF system back from a solar developer under the PPA. The price may vary among different solar developers but is negotiable [23]. Through a PPA, MnDOT only needs to take care of the costs of the structural snow fences and thus has very minimum initial costs for a PVSF project, which, hence, makes the parameters related to the structural snow fences more predominant, such as PVSF 3 and 10. PVSF 12 represents the current price of the electrical power taken from the electrical grid and paid by MnDOT to a utility company. The difference between the PPA price (PVSF 9) and the utility price (PVSF 12) determines whether an annual cost or benefit will be incurred to the cash flows when purchasing the same amount of electrical power to support its facilities of MnDOT near highways, such as rest areas. If the PPA price is higher than the current utility price paid by MnDOT, it would cause an increase in the annual costs in the cash flow calculation and thus result in a reduction in cost-effectiveness when compared with a case where the PPA price is lower than the current utility one, which would help MnDOT to save money on an annual basis.

As shown in Table 10, similar critical parameters were also identified for IRR in the case of Category 2, i.e., PVSF 3, 9, 10, and 12. Due to the fact that PVSF 11 will not change in IRR calculations (IRR is the discount rate at which the NPV is zero), PVSF 4, i.e., Land Costs of Snow Fences, thus becomes predominant in a PPA-based PVSF project. Structural snow fences are typically installed on farmland along highways and MnDOT has two methods to obtain permission from a private landowner to install snow fences: (1) purchasing the property through an easement or (2) leasing farmland. This study assumes the farmland for the snow fences is leased with the rental cost of USD 3.28/linear meter/year (for a property with a width of 7.62 m), which is a more popular method for MnDOT in a snow fence project. All the PVSF 3, 10, and 4 are the parameters associated with the structural snow fences, and thus it is clear that from the perspective of MnDOT, the cost and benefit of the snow fence itself play a significant role in determining the profitability of a PVSF project through a PPA, since most of the costs related to the PV system would be covered by a solar developer.

After examining the most sensitive parameters, it is important to look at the most insensitive parameters to the analysis results in terms of NPV and IRR. For Category 1, they are RECs (PVSF 7), O&M costs of snow fences (PVSF 5), annual costs of the PV system (PVSF 2), and land costs of snow fences (PVSF 4). Please note, PVSF 9 and 12 (in estimating NPV and IRR), as well as PVSF 1 (in estimating IRR), have zero % difference (Table 7), since they were not used in these cash flow calculations of Category 1 and thus were not considered as insensitive parameters. PVSF 13 (Orientation) is a parameter that has a medium impact on the cost-effectiveness of a PVSF project when MnDOT invests in and owns the PV system. For Category 2, the most insensitive parameters for project NPV and IRR are Orientation (PVSF 13) and O&M costs of snow fences (PVSF 5), while other parameters that have a different % of zero, as listed in Table 8, were not used in the financial analysis model and thus were not considered. These identified insensitive parameters play a less important role than those critical parameters identified, and thus less attention would be paid to them by decision makers depending on the financial mechanism between Category 1 and 2.

Considering the uncertainties of the benefit parameters including PVSF 6, 7, 8, 9, 10, and 12, which may vary depending on different governments and/or policies, an additional sensitivity analysis was carried out by eliminating their impacts on the analysis results. This means in the financial analysis model, PVSF 6, 7, 8, and 10 were set to be zero (USD 0 or 0%), while PVSF 9 and 12 were set to be identical, i.e., no difference exists between the PPA price and the electricity utility price paid by MnDOT.

Table 11. Sensitivity results of critical parameters for costs (Category 1).

PVSF Parameters	NPV
	Mean	SD	Range	Diff%	Rank
1	USD (635,416.50)	USD 300,467.97	USD 737,621.45	72.74%	1
2	USD (640,740.94)	USD 40,411.48	USD 91,328.48	13.37%	6
3	USD (633,058.65)	USD 581,329.56	USD 1,429,473.10	39.38%	2
4	USD (628,053.02)	USD 36,031.18	USD 91,941.42	13.47%	5
5	USD (630,844.70)	USD 20,347.92	USD 52,239.44	7.88%	7
6	-	-	-	-	-
7	-	-	-	-	-
8	-	-	-	-	-
9	USD (636,695.51)	0	0	0.00%	8
10	-	-	-	-	-
11	USD (623,756.29)	USD 122,491.34	USD 297,322.84	38.78%	3
12	USD (636,695.51)	0	0	0.00%	8
13	USD (710,085.91)	USD 100,502.86	USD 185,369.63	22.55%	4

shows the Category 1 NPV sensitivity results of the critical parameters for costs (without considering the benefits), and

Table 12. Critical parameters for costs (Category 1).

Rank	NPV
	Parameter Index	Parameters
Top 1	PVSF-1	Total Capital Costs of the PV system
Top 2	PVSF-3	Total Capital Costs of Snow Fences
Top 3	PVSF-11	Real discount rate
Top 4	PVSF-13	Orientation
Top 5	PVSF-4	Land Cost of Snow Fences

shows the corresponding top critical parameters identified. The sensitivity results for Category 2 are shown in

Table 13. Sensitivity results of critical parameters for costs (Category 2).

PVSF Parameters	NPV
	Mean	SD	Range	Diff%	Rank
1	USD (276,230.70)	0	0	0.00%	6
2	USD (276,230.70)	0	0	0.00%	6
3	USD (272,593.84)	USD 126,311.90	USD 312,164.16	72.21%	1
4	USD (267,588.21)	USD 36,031.18	USD 91,941.42	28.54%	2
5	USD (270,379.88)	USD 20,347.92	USD 52,239.44	17.28%	3
6	-	-	-	-	-
7	-	-	-	-	-
8	-	-	-	-	-
9	USD (389,226.29)	0	0	0.00%	6
10	-	-	-	-	-
11	USD (275,358.87)	USD 9320.18	USD 22,639.69	7.91%	5
12	USD (389,226.29)	0	0	0.00%	7
13	USD (295,427.29)	USD 26,288.34	USD 48,486.78	14.93%	4

and

Table 14. Critical parameters for costs (Category 2).

Rank	NPV
	Parameter Index	Parameters
Top 1	PVSF-3	Total Capital Costs of Snow Fences
Top 2	PVSF-4	Land Cost of Snow Fences
Top 3	PVSF-5	O&M costs of snow fences
Top 4	PVSF-13	Orientation
Top 5	PVSF-11	Real discount rate

.

As shown in Table 11 and Table 13, the mean NPVs are all negative, because only costs were considered in the financial analysis model, which is also the reason why IRRs are not shown. Without considering any benefits, IRR has no real meaning. For Category 1, unlike the results with benefits (Table 9), the capital costs of the PV system (PVSF 1) and snow fences (PVSF 3) play a significant role, compared to the real discount rate (PVSF 11). For Category 2, without considering any benefits, the real discount rate becomes less important compared to the costs associated with the snow fences, such as PVSF 3, 4, and 5 in the financial analysis.

Table 15. Other considerations/factors for a PVSF project.

Rules, Regulations, and Policies	Electrical permits are typically required for solar PV systems [41]. Work on large plants must comply with the Minnesota State Building Code (Minnesota Statutes section 216E.10). Large electric-power-generating plants are defined as 50,000 KW or greater in capacity by Minnesota Statute 216E. They can include solar PV-collection sites in multiple locations if all sites are constructed within 12 months of each other, and are owned and funded as a single project. These cases do not include the electric-power generator plants that are owned and operated by a public utility [41]. Utility Accommodation: The FHWA has determined that using a highway ROW to accommodate public-utility facilities is in the public’s interest. To the extent that these facilities serve “the public,” they can be accommodated under the DOT’s approved Utility Accommodation Policy (UAP). If utilizing such facilities serves a private or proprietary interest, the use would have to be approved under the ROW-use agreement requirements (23 CFR 710 Subpart D). Thus, the distinction between public or private use, which is defined by state statute, determines which regulations apply [42]. Control of Access to the Interstate: The FHWA retains all approval rights to control access to the interstate system (23 CFR 620.203h). A DOT is required to obtain the FHWA’s written approval when access to the interstate system is added or modified. Both temporary and permanent modification of access control for transportation and non-transportation purposes requires FHWA approval [42]. Use and Occupancy Permit: Renewable-energy facilities on the highway ROW require a Use and Occupancy Agreement (23 CFR 645.213). The permit must reference the state’s DOT standards and include the following information: a description of the facility’s type, size, and location; an adequate drawing that shows the existing/proposed facility, ROW lines, control of access limits, and approved access points; the extent of liability and responsibilities for future adjustments; and the action to be taken for non-compliance [42]. National Environmental Policy Act (NEPA): Under federal law, the NEPA applies to any proposed action or transportation project where federal funds or assistance will be used at some phase of project development or where federal permits or approvals are required. FHWA approval of ROW Use Agreements or access control constitutes a federal action, thus requiring that the NEPA procedures be followed [42]. Subpart B: Accommodation of Utilities of the Electronic Code of Federal Regulations, Title 23, Chapter I, Subchapter G, Part 645 (Office of the Federal Register (OFR) and the Government Publishing Office). Prohibition of Commercial Establishments in the ROW [43] When the research is implemented, the legal team of MnDOT may need to confirm that selling electricity to a utility is not considered commercial activity on the ROW.
Insurance	Most large PV systems require liability and property insurance, and many developers may opt to add policies such as environmental-risk insurance [44]. General liability covers policyholders for death or injury to persons or for damage to property owned by third parties. Property-risk insurance covers “damage to or loss of policyholders” property. While the manufacturer’s warranty will provide some limited defect coverage, the system owner usually purchases property insurance to protect against risks not covered by the warranty or to extend the coverage period. Environmental damage coverage indemnifies system owners of the risk of either environmental damage inflicted by their development or pre-existing damage on the development site. Developers estimate the annual insurance cost to be around 0.25% of the project’s total installation cost; the amount could be as high as 0.5% annually in areas where extreme weather events are likely [44].
Interconnection	Interconnection Standards: “Interconnection” refers to the physical linking of an energy generator to the larger electric grid. Assuming that a renewable energy system will be connected to the electricity grid, local electric utilities manage that interconnection [42]. Information from Xcel Energy: Xcel Energy is required to follow Minnesota’s statewide Distributed Energy Resources Interconnection Process (MN DIP) for all solar interconnections [45]
Tax Credits/Incentives	The investment tax credit (ITC) and the production tax credit (PTC) both provide tax relief to a renewable-energy developer in an amount that depends on the program available, which for solar projects has been the ITC. Under the ITC, the IRS approves a tax credit that is equal to 30% of the project’s total cost. The tax credit can be claimed when filing federal income taxes subsequent to the project going into service. The tax credit’s value may be sold to other private entities which have a tax interest and see an economic advantage to partnering on the project. With the tax credit, the private entity never receives cash to help pay for a project but rather receives a credit on the payments due, thereby producing a net positive on the company’s balance sheet [30]. A source to examine when searching for policies and incentives by state is DSIRE [40].
Sales, Income, or Property Taxes	Solar equipment has an exemption from state sales tax; the tax exemption would need to be vetted/confirmed whether it applies to MnDOT and its contractors, as well as the general public [46]. Farmland, if used to place snow fences, will be considered as Commercial/Industrial/Public Utility use, instead of its original intended use, when landowners pay property taxes if the land is leased to MnDOT. No property taxes will be paid if MnDOT purchases and owns farmland to install snow fences.
Other considerations about PPA	The PPA contract term is flexible, but usually ranges from 15 to 25 years. Renewable Energy Certificates (RECs) are typically owned by the developer/investor. The net metering cap is 40 kW, and a project with 40 kW or greater will have a lower energy selling price (around USD 0.02~0.03/kWh). PPA price varies project by project but is negotiable. The location of the project matters depending on how far the system is to connect to the grid. Newly developed panels and/or system (for PVSF) will increase the uncertainties, and the solar developer will work with MnDOT to evaluate the project first before proceeding any further.

provides the possible effects of rules, regulations, policies, insurance, the utility company’s interconnection requirements, and/or tax credits/incentives on the implementation of a PVSF project for MnDOT, which would also be good references for other state DOTs.

5. Conclusions

PV Snow Fences (PVSF) is a system that integrates structural snow fences with PV panels. PVSF combines the benefits of both structural snow fences and PV panels that can be used not only in summer to generate renewable energy but also in winter to eliminate blowing and drifting snow on highways, which contributes to environmental sustainability and the continuous development of the structural snow fence and PV technologies. This paper demonstrates the results of a financial analysis for a PVSF project to be implemented in Minnesota by MnDOT, especially the results of a sensitivity analysis, including identifying critical parameters, as well as how these parameters would affect the cost-effectiveness of the PVSF project. This study is expected to provide references to MnDOT and/or other DOTs for the development and implementation of PV snow fence projects in the US. It can be considered as the preliminary stage of a PVSF project for decision makers in deciding whether additional information is needed to narrow the range of estimates for some parameters or in identifying the critical parameters to be investigated in detail in a more formal risk and uncertainty analysis [13]. The conclusions condensed from the financial analysis results are summarized below.

➢

For Category 1, where the PV system on structural snow fences is owned by an organization, e.g., MnDOT in this project, the most sensitive parameter in determining project NPV is the real discount rate. A relatively low real discount rate caused by COVID 19, if its impact would last for years with such a low rate, would be helpful for a PVSF project to be more cost-effective compared with a project with a higher rate.

➢

For Category 1, the most sensitive parameter in determining project IRR is the cost of the PV system, and according to the current PV market [36], it is believed that the cost of a PV system would be reduced year after year, which would contribute to the cost-effectiveness of a PVSF project.

➢

For Category 2, where the PV system is owned by a solar developer via a PPA rather than MnDOT, the most sensitive parameter in determining project NPV is still the real discount rate, which indicates that the real discount rate plays the most significant role in a PVSF project. Again, PPA stands for Power Purchase Agreement, where an organization, e.g., MnDOT, accrues the benefits of solar power without owning the system by buying the power generated at a negotiated rate from a solar developer, who procures, builds, operates, and maintains a solar system [23].

➢

For Category 2, the most sensitive parameter in determining project IRR is the cost of structural snow fences, since the cost of the PV system would be covered by a solar developer through a PPA. This indicates that the cost-effectiveness of a PPA-based PVSF project is predominantly determined by the costs and benefits associated with the structural snow fences rather than the PV system.

➢

According to the top 5 sensitive parameters identified (Table 9 and Table 10) in determining project NPV and IRR for both of Category 1 and 2, in addition to the real discount rate (PVSF 11), the cost-effectiveness of a PVSF project via a PPA or not is mainly dependent on the costs of either the PV system or the snow fences or both, such as PVSF 1, 3, and/or 4, as well as the benefits of either the PV system or the snow fences or both, such as PVSF 8, 9, 10, and/or 12.

➢

For Category 1, the most insensitive parameters to the analysis results in terms of NPV and IRR are RECs, O&M costs of snow fences, annual costs of the PV system, and land costs of snow fences, while for Category 2, the most insensitive parameters are Orientation and O&M costs of snow fences. These identified insensitive parameters play a less important role than those critical parameters identified in the cost-effectiveness of a PVSF project.

➢

The real discount rate has less significance in the financial analysis compared with the cost parameters for the PV system and/or snow fences, such as PVSF 1 or 3, when project benefits were not considered in the analysis due to their uncertainties, which would vary depending on the different governments and/or policies.

➢

It is expected that MnDOT or other DOTs, who would develop and implement a PVSF project, pay more attention to the critical parameters identified in this study depending on who will own the PV system (Category 1 or 2), since their changes would have more significant effects on the financial analysis output (NPV and IRR).

This paper summarizes the sensitivity analysis results of a financial analysis for a PVSF project to be implemented in Minnesota. Other technical findings of the MnDOT-funded project, including the structural design of the PVSF, the development of the electrical system, the cost-benefit tradeoff analysis, the impacts of different factors, such as weather, orientation, etc., on the financial analysis results, and the evaluation of the snow trapping behavior by using PV panels, can be found from the MnDOT project report [47] and would be reported separately. Additionally, the current study assumes the parameters studied are independent, and the possible correlations between them would be identified in future studies, along with their uncertainties in the form of probability distributions.