Cost Assessment of Centralizing Swine Manure and Corn Stover Co-Digestion Systems

Abstract

Livestock in the state of Iowa, United States (US) produce over 50 × 10⁶ Mg of wet-basis manure yearly. Biogas production from manure’s anaerobic digestion (AD) can reduce greenhouse gas emissions, control odors, and provide renewable energy. Despite these benefits, AD is rarely deployed at swine farms in Iowa. In this work, we explore the economics of AD systems in Iowa to evaluate reasons for low deployment and explore the production cost impacts of biogas cleaning and injection into the natural gas grid, amending manure with biomass, and centralizing digesters across multiple farms. This work presents a static, spreadsheet-based technoeconomic model that embodies literature-based estimates of key system technical parameters, costs, and transportation fuel incentives and permits the examination of various scenarios. Key findings include that under the model assumptions, distributed, farm-scale digesters are not competitive with average natural gas prices in Iowa. A centralized production scenario can be competitive, provided that programs such as the low-carbon fuel standard (LCFS) and the renewable fuel standard (RFS) have sufficiently high credit values.

1. Introduction

Anaerobic digestion (AD) is a process by which microbes break down organic materials in an oxygen-free environment, resulting in two output streams: a stabilized digestate typically suitable for land application and an energy-rich gas referred to as biogas. Anaerobic digestion feedstocks include industrial and municipal wastewater, manure, crop residues, and food waste [1]. Biogas is composed mainly of greenhouse gases, namely methane and carbon dioxide. The AD process captures the methane and displaces fossil fuels when it is utilized as an energy source [1]. In addition to providing a renewable energy source that is independent of short-term weather fluctuations (i.e., wind, sunlight), AD of animal manure offers other benefits, including reductions in odor from, and pathogen populations in, the manure slurry, as well as reduced overall methane emissions [1,2]. Digester effluent (i.e., digestate) can be applied to cropland to improve soil health and provide critical macronutrients [3].

In 2020, manure-based anaerobic digesters in the United States (US) were estimated to have reduced greenhouse gas emissions by 5 × 10⁶ Mg CO_2eq and generated the energy equivalent of 15 × 10⁶ GJ [4]. In 2021, there were 273 operational manure-based anaerobic digesters in the US, with the majority (81%) processing dairy manure and 16% processing swine manure [4]. According to AgSTAR [5], swine-manure-based digestion systems can be profitable at farm sizes greater than 2000 head. Iowa has more swine farms (~2200 farms) that fit this category than any other US state [5]. The opportunity to achieve economies of scale on Iowa swine farms has improved, as most of Iowa’s hogs (69%) are now produced on farms greater than 5000 head and the common farm size in the state is between 2000–4999 head [6]. Economies of scale appear to be necessary for current AD technologies.

Multiple investigators have attempted to understand the costs of farm-based energy from swine AD with a wide range of results. Bhatt and Tao [7] modeled costs for swine manure digestion ranging from flows of 1–227 Mg wet manure per day, finding estimated energy costs of USD 20–USD 130 per GJ, significantly higher than the cost of natural gas in the US, which averaged USD 3.54 per GJ from 2016–2020 (Citygate: [8,9]). A prior study by Beddoes et al. [10] reported swine manure AD followed by electricity generation to provide electricity at USD 20–USD 30 per GJ, which is far more expensive than current renewable electricity options, such as photovoltaic or wind [10,11]. When considering only the biogas production portion of these systems, Beddoes et al. [10] found that the cost of energy production ranged from USD 3–USD 6 GJ⁻¹, with the lowest value (USD 3 GJ⁻¹) achieved by a single covered lagoon digester. Faulhaber et al. [12] explored the impact of digester size on AD energy costs and defined a dimensionless methane cost ratio by dividing the biogas energy cost (USD GJ⁻¹) by the retail price of natural gas (USD GJ⁻¹) [12]. They showed that under baseline assumptions, even large (1000-head dairy) systems had methane cost ratios greater than unity (meaning they were more expensive than natural gas), but that carbon credits and low-interest loans could make these systems more economically competitive.

In addition to these results, the failure rate of US farm-based digesters is relatively high, e.g., in 1998, this rate was approximately 50% [13]. According to more recent AgSTAR data, failure rates have decreased to approximately 23% [14]. Unfavorable economic factors and system operations, which require management, time, and skill combined to make AD economically challenging. Still, centralized digesters handling the manure of multiple farms may allow a dedicated operations staff to manage day-to-day maintenance tasks, reducing the management burden on farmers and potentially reducing digester abandonment [15].

The majority of anaerobic digester projects since 2000 have been combined heat and power systems, but in recent years, projects focused on producing renewable natural gas (RNG) for use in vehicles have become popular [4]. The number of RNG projects that upgrade biogas for pipeline injection in the US increased by 85% from 2019–2020 [16]. These RNG projects have greater capital costs due to the need for equipment to upgrade the biogas for pipeline injection. However, when marketed as a transportation fuel, RNG production allows methane to qualify for renewable identification numbers (RINs) and the low-carbon fuel standard (LCFS) creating an additional revenue source for the biogas producer.

The economic viability of AD is challenging. ADs are bioreactors with relatively low volumetric productivities, i.e., <0.1 g biogas L⁻¹ h⁻¹ [17,18] compared to 2 g ethanol L⁻¹ h⁻¹ or greater for ethanol fuel production fermenters [19]. So, although AD reactors are rather simple and low cost to make, the low volumetric productivities remain challenging. One approach to overcoming this economic challenge is to increase the volumetric productivity of the digester by increasing its feed rate, specifically by adding other carbonaceous feedstocks—typically agricultural or food residues—to the influent [20,21]. This approach is termed co-digestion and is one of three approaches used to make AD more cost effective. Other approaches include decreasing the unit cost of digestion by building one digester for multiple farms and cleaning and injecting the resulting methane into a pipeline to realize carbon credits available in certain markets.

In this work, a static spreadsheet-based technoeconomic model is built to explore the costs of AD for large-scale swine farms typical of Iowa, aiming to clarify the challenges and opportunities for greater AD deployment. We expand beyond previously published models by considering scenarios with RIN and LCFS credits to reflect the current carbon credit climate and including the costs of upgrading biogas to RNG to qualify for those credits. To explore the possibilities of leveraging economies of scale, we model both one-digester-per-farm systems and joint (among five farms) systems. Additionally, we examine the impact of co-digesting corn stover with swine manure to increase the energy production of the digester because corn stover is a highly available crop residue in the US Midwestern region [22]. Finally, we consider the impact of distance to an injection point on the system’s economic viability.

2. Materials and Methods

2.1. Scenario Description

The cost of producing biogas was calculated for four different scenarios (represented as S1 through S4), reflecting typical Midwest agriculture in 2020. In each scenario, costs were calculated based on farm sizes of 4800 pigs, representative of a common farm with two 2400-head barns [23,24]. S1 consists of a single digester, receiving deep-pit effluent. Corn stover is added until 12% solids is reached, but no water is added. Wet digestion processes operate between 5–20% total solids (TS) [25]. The biogas is cleaned and injected into the natural gas grid on site. In the second scenario (S2), a volume of water equivalent to manure volume is assumed to be added to the digester to permit higher corn stover additions within the 12% solids limit. This leads to higher biogas production but requires a larger digester volume. This is a critically important aspect of this model. S2 and subsequent scenarios allow additional carbon feedstock into the digester and require a larger digester volume. In S3, the benefits of centralized biogas cleaning are explored by assuming that five decentralized digesters share the cost of a single centralized biogas cleaning (also referred to as upgrading) and injection point, as investigated previously by Hengeveld et al. [26]. The assumption is that farms are 4 km from the centralized point, which is realistic for the concentrated production areas of the region being modeled [27,28]. In S3, biogas must be transported, but each farm uses its own manure, stover, and water. Finally, in S4, we examine the impact of leveraging economies of scale by centralizing digestion, cleaning, and injection among five farms. We account for the cost of these five farms hauling their manure to the central digester where the biogas is produced, cleaned, and injected. A summary of the differences between the scenarios is shown in

Table 1. Summary of Modeled Scenarios: Total solids (TS); Anaerobic Digester (AD).

Parameter	S1	S2	S3	S4
Farm size	4800 pigs	4800 pigs	4800 pigs	5 farms of 4800 pigs
Location of digester	On-farm	On-farm	On-farm	Centralized location: 2.5 miles from 5 farms
Water addition	None	Equivalent to manure volume	Equivalent to manure volume	Equivalent to manure volume
Corn stover	Added to 12% AD TS	Added to 12% AD TS	Added to 12% AD TS	Added to 12% AD TS
Biogas upgrading	On-farm	On-farm	Centralized location: 2.5 miles from 5 farms	At centralized AD site

. To help readers keep track of the scenarios, we use the following shorthand descriptions: S1: single digester, stover amended; S2–S1: w/water and stover amended; S3–S2: w/centralized upgrading; and S4: centralized digester/upgrading, water and stover amended.

2.2. Digester Input

In each scenario, it is assumed that all the manure produced is used in biogas production. The base manure production assumption is 4.5 L⁻¹ day⁻¹ head⁻¹ [29]. The digesters are assumed to operate 24 h per day for 340 days each year. The assumed methane production for swine slurry is 350 mL g⁻¹ volatile solids (VS) [21,30,31]. We assume a TS content for the manure of 8%, a VS content of 6%, and a density identical to water (1000 kg m⁻³) [32]. The cost of manure was calculated based on its nitrogen content, 6 kg m⁻³ (0.05 lb gal⁻¹) [29], and assuming 95% of the nitrogen to be plant-available nitrogen [33]. We assumed that nitrogen has a value of USD 1.06 kg⁻¹ [34]. This cost is applied to all scenarios because we assume that a separate entity from the farm runs the digesters. Manure transportation costs were calculated according to Equation (1), where d is the two-way transportation distance in km and C_MT is the cost of manure transportation in 2019 USD Mg⁻¹. In S4, this calculation amounted to USD 8.40 Mg⁻¹ (2019 USD), assuming a 4 km transportation distance for each farm [35]. On-farm manure transportation costs are not considered because those operations occur even in non-AD cases. The only manure transportation costs considered in this work are those required for off-farm transport, as in S4.

$$C_{MT} \left(\$ {\mathrm{Mg}}^{-1}\right)=\$0.27d+\$2.88$$

(1)

According to Moody et al. [36], the methane production of untreated corn stover is assumed to be 180 mL g⁻¹ VS, and the TS content and VS content of wet corn stover are assumed to be 90.3% and 84.1%, respectively. The lignin content of corn stover inhibits its biodegradability and, therefore, its methane production [21]. Pre-treatment of corn stover can enhance methane production, but this analysis focuses on untreated corn stover. The assumed methane production is lower than could be achieved if the corn stover was pre-treated because untreated corn stover has low biodegradability [37,38]. Corn stover costs vary in estimates and models from USD 20–100 Mg⁻¹ [39,40,41,42,43,44,45]. In this study, a value of USD 51 Mg⁻¹ was selected (USD 35 bale⁻¹ [45]). Given the large variability and the importance of this assumption, the sensitivity of the S1 and S4 configurations to the price of corn stover is explored. The price is assumed to include the transportation of the corn stover to the digesters. The digesters are assumed to achieve 75% of the methane production assumptions from both inputs. In each scenario, the mass of corn stover added to the digesters is calculated to meet a 12% solids content after the addition of manure and water. Corn stover accounts for 37% of the input solids in S1 and 70% of the input solids in S2–S4.

There are two critical land areas in this system: the land required to receive the nutrients in the digestate safely and the land required to provide stover as an energy amendment to the digester. We compare these two land requirements as follows. We assume corn stover is produced at a rate of approximately 11 Mg ha⁻¹ (dry matter; 5 US ton ac⁻¹ [46]). Stover harvest recommendations range from 30–50% based on tillage practices at the farm [47]. Because we assume digestate will be land applied, we assume an aggressive harvesting rate of 60% or 6.7 Mg ha⁻¹. This results in a land requirement for biomass demand, ranging from 56 ha yr⁻¹ (140 ac yr⁻¹) in S1 to 1100 ha yr⁻¹ (2760 ac yr⁻¹) in S4. We again assume the manure’s nitrogen content to be 0.006 Mg m⁻³ and assume the corn stover’s nitrogen content to be 6 kg N per Mg corn stover [48]. We assume digestate is applied at a rate of 168 kg N ha⁻¹ (150 lb ac⁻¹). This results in a land application requirement of 275 ha in S1, 315 ha in S2–S3, and 1580 ha in S4. Therefore, the land needed for biomass harvest is 71% of the digestate application land requirement at its maximum demand.

Table 2. Biomass Use/Land Application Balance. S1 = single digester, stover amended; S2 = S1 w/water and stover amended; S3 = S2 w/centralized upgrading; S4 = centralized digester/upgrading, water and stover amended.

	Unit	S1	S2	S3	S4
Corn stover demand	Mg yr−1	375	1500	1500	7500
Corn stover harvest rate	Mg ha−1	6.7	6.7	6.7	6.7
Corn stover harvest land requirement	ha yr−1	56	223	223	1116
Digestate nitrogen content	Mg yr−1	46.3	53	53	265
Land application requirement	ha yr−1	275	315	315	1576
Portion of land application area needed for stover harvest		20%	71%	71%	71%

shows the details of this calculation. Harvesting corn stover at a 50% rate (5.6 Mg ha⁻¹) would result in corn stover harvest on 85% of the land needed for digestate application in S2–S4. At a 30% harvest rate, the harvest and application land areas would not be in balance (corn harvest on 142% of digestate application land). Although corn stover cost is treated as a constant in this analysis, harvesting at a lower rate and a larger land area would likely increase costs associated with using corn stover as an AD feedstock.

2.3. Biogas Production

Total possible methane production for the digester in each scenario was computed based on digester inputs and conversion rates as described above. After applying the 75% digester efficiency assumption, the methane production of the digester in S1 was 209 m³ kg VS⁻¹ and 168 m³ kg VS⁻¹ in the other three scenarios, in line with values reported in manure and biomass co-digestion studies [20,21,49,50]. Methane was converted to biogas by assuming the biogas is 60% methane by volume. The volume fraction of methane in biogas in this type of system varies, but reported values reported range from 50 to 70% [39,51,52,53]. The annual biogas production ranged from approximately 0.23 to 2.4 × 10⁶ m³, depending on scenario. The annual energy production in each scenario was calculated by assuming the energy density of methane is 37.6 MJ m⁻³ (1010 BTU ft⁻³; [54]) and by applying a biogas cleaning efficiency of 97%, meaning 97% of methane is available after the biogas upgrading process [55].

2.4. Digestate Production

Solid and liquid digestate production was calculated as the remaining mass after biogas was produced. By assuming the biogas produced is 60% methane, 6% water vapor, and 34% carbon dioxide by volume, the total mass of biogas was calculated. The mass of solid digestate was calculated by subtracting the mass of methane and carbon dioxide from the input solids. The mass of liquid digestate was calculated by subtracting the mass of biogas and solid digestate from the total digester input.

We assumed a recycle rate of 50% [56,57] (50% of the liquid at the solid separator is returned to the digester, 50% of the effluent sent to liquid storage). After the digester reaches the steady state, both the solid and liquid portions of the digestate can be land-applied and have value as a fertilizer. The value of this fertilizer was calculated based on the manure and corn stover’s nitrogen contents, assuming N is conserved in digestion. Again, we assumed nitrogen has a value of USD 1 kg⁻¹, and the plant-available nitrogen contents of the digester inputs were 95% for the swine manure [33] and 50% for the corn stover.

2.5. Capital Costs

Digester volume for our scenarios was calculated according to Equation (2), where HRT is the hydraulic retention time in days; S is herd size; M is daily manure production in m³ d⁻¹ head⁻¹; CS_vol is the volume of corn stover added to the digester; and W_vol is the volume of water added to the digester.

$$V=HRT\left(S\times M+C{S}_{vol}+W_{vol}\right)$$

(2)

Capital costs for plug flow digesters were presented by Faulhaber et al. [12] in terms of dairy farm herd size as shown in Equation (3), where C_D is the capital cost of the digester and S is the herd size in terms of the number of cows. Equation (4) was used to relate Faulhaber et al.’s [12] herd size to the digester volume. In Equation (4), V is the digester volume in m³, HRT is the hydraulic retention time in days, and M is the daily manure production in m³ d⁻¹ head⁻¹. Equation (4) enables the conversion from herd size to digester volume for the use of the Faulhaber et al.’s [12] equation in our work. According to the ASABE Manure Production and Characteristics Standard [32], dairy cows produce manure at a rate of 0.068 m³ d⁻¹ head⁻¹. Assuming a hydraulic retention time of 30 days, combining Equations (3) and (4), and converting from 2012 USD to 2019 USD, we can calculate the capital cost of the digesters according to Equation (5). In the final cost calculation, the manure production value for swine mentioned previously (4.5 L⁻¹ day⁻¹ head⁻¹) was used to calculate digester volume according to Equation (2).

$$C_D\left(\$\right)=\$13,575\times S^{0.59}$$

(3)

$$S=\frac{V}{M\times HRT}$$

(4)

$$C_D\left(\$\right)=\$9900\times V^{0.59}$$

(5)

The capital cost of biogas upgrading and RNG injection equipment was adapted from the report by Williams et al. [58]. Annual upgrading, injection, piping, and compression capital costs were adjusted from 2015 USD to 2019 USD, and the California multiplier was excluded. These capital costs were divided by the methane flows given in the report for a unit capital cost in USD GJ⁻¹. The unit capital costs were plotted against the methane flows as shown in Figure 1. The unit capital cost for these components in S1–S4 was calculated using the power trend line from Figure 1 and the methane flows in each scenario. This resulted in unit capital costs of upgrading and injection equipment ranging from USD 9.54 GJ⁻¹ in S3 and S4 to USD 26.80 in S1.

Pipeline costs were calculated in S1–S3. In S1 and S2, the digesters are assumed to be located 500 m from the natural gas grid injection point [55]. In these cases, we assume a relatively short distance from the digester to the injection point because we assume a farm would not undertake a new RNG project when located very far from an injection point. In S3, biogas is transported 4 km (2.5 miles) from the digester to the centralized cleaning and injection point. Hengeveld et al. [26] presented installed pipeline capital costs in the Netherlands ranging from USD 50,000 km⁻¹ (USD 80,000 mile⁻¹) to USD 373,000 km⁻¹ (USD 600,000 mile⁻¹), depending on the ease of installation and pipeline diameter (converted from 2014 EUR to 2019 USD) [26]. The EPA Landfill Gas Energy Cost Model [59] estimates that pipeline costs for RNG projects are USD 600,000 total for pipelines less than 1.6 km (1 mile) and USD 0.62 M km⁻¹ (USD 1 M mile⁻¹) for pipelines one mile or longer. In this model, we utilize the EPA figure but explore how the Hengeveld assumptions influence costs. For S1 and S2, the pipeline capital to transport RNG to the injection point is USD 600,000 total because the distance is less than 1 mile. In S3, the distance was assumed to be 2.5 miles, which corresponds to a cost of USD 2.5 M. We assume the centralized biogas cleaning point in S3 and the centralized digester in S4 is located directly on an injection point (distance = 0 m), so no additional pipeline cost is considered. In S3, additional costs of biogas compression, H₂S removal, and dewatering were calculated based on the findings of Hengeveld et al. [26] and converted to 2019 USD.

Capital costs for solids separation for digestate were calculated as screw press separators according to Møller et al. [60], assuming each of the farm-scale digesters in S1–S3 would require one separator and S4 would require five. Storage for digestate was calculated as one year of manure volume stored in a covered lagoon to account for the timing of land application. Lagoon cost was calculated as USD 7.34 m⁻³ (USD 0.028 gal⁻¹) [61]. Assuming a depth of 15 ft, the cost of the cover was estimated as USD 11 m⁻² (USD 1 ft⁻²) of the surface area [62]. All of the capital costs, excluding the upgrading and injection capital, were summed and an annual capital cost was computed assuming a 15-year lifespan at 7% interest. This was converted to a unit capital cost based on the energy production (USD GJ⁻¹), then added to the capital cost of upgrading and injection.

Table 3. Capital Cost Calculations in 2019 United States Dollars (USD).

Parameter	Calculation	Comment
AD cost	$C_D=\$9900\times V^{0.59}$	$C_D=$ capital cost of digester $V$= volume of digester, m3 Adapted from Faulhaber et al. [12]
S1 and S2 pipeline cost	USD 600,000	Pipeline for RNG to injection point: EPA [59] suggestion for pipelines shorter than 1.6 km
S3 pipeline cost	USD 620,000 km−1	Pipeline for raw biogas to centralized upgrading and injection point: EPA [59] suggestion for pipelines longer than 1.6 km
Lagoon	USD 7.34 m−3 (lagoon) USD 11 m−2 (cover)	Storage calculated for one year of manure volume
Upgrading and injection	$C_{UI}=22.794\times x^{-0.469}$	x$=\mathrm{methane} \mathrm{flow} \mathrm{to} \mathrm{the} \mathrm{equipment}, \mathrm{GJ} {\mathrm{hr}}^{-1}$ $C_{UI}$= Unit capital cost of the upgrading and injection equipment, USD GJ−1 Adapted from Williams et al. [58]
S3 biogas compressor	$C_{BC}=\$173.5\times 0.23B$	$C_{BC} =$ Investment cost of biogas compressor, USD 1000 s $B$= hourly biogas flow, m3 hr−1 Adapted from Hengeveld et al. [26]
S3 biogas dewatering and H2S removal	USD 7800	Adapted from Hengeveld et al. [26]
Screw press cost	USD 18,890	Adapted from Møller et al. [60]: assumed one needed in S1–S3, five needed in S4
Interest rate	7%	Assumption
Project lifetime	15 years	Assumption

summarizes the capital cost components for all scenarios.

2.6. Operating Costs

Operating costs for labor, maintenance, biogas upgrading and injection, energy, and solids handling were calculated. Labor costs are based on Aui and Wright [39], which included salaries for a plant manager and two yard employees for a total annual commitment of USD 190,000 for a digester producing 8300 m³ biogas per day. The labor cost was normalized based on biogas production, converted to 2019 USD, and scaled linearly for each scenario. Annual operating and maintenance cost of upgrading and injection were calculated as USD 42,400 m⁻³ min⁻¹ (USD 1200 scfm⁻¹) capacity [59]. An additional maintenance cost was applied as 10% of the annual capital cost, excluding the cost of upgrading and injection. The digester’s energy use was assumed to be 1.24 MJ m⁻³ biogas and the screw press energy demand was assumed to be 0.53 kWh Mg⁻¹ [26,60]. The average cost of industrial electricity in Iowa is USD 0.0689 kWh⁻¹ [63]. An additional cost for solids handling was included as USD 5.00 Mg⁻¹ of solid digestate [39]. Other operating costs covered previously include manure and corn stover costs and the cost of manure transportation.

Table 4. Operating Costs in 2019 USD.

Parameter	Value	Note
Manure cost	USD 86.40 m−3	Based on the plant-available nitrogen content of the manure, discussed in Section 2.2
Corn stover cost	USD 51 Mg−1	From Edwards [45]; range discussed in Section 2.2 and explored in Section 3.5
S4 manure transportation	$C_{MT} \left(\$ {\mathrm{Mg}}^{-1}\right)=\$0.27d+\$2.88$	d = two-way transportation distance, km Adapted from Ghafoori et al. [35]
Labor	USD 0.08 m−3 biogas production	Adapted from Aui and Wright [39]
Upgrading and injection, O and M	USD 42,400 m−3 min−1	Annual cost based on capacity; adapted from EPA [59]
Solid digestate handling	USD 5 Mg−1	Aui and Wright [39]
Digester energy demand	1.234 Mg m−3 biogas	Hengeveld et al. [26]
Screw press energy demand	0.53 kWh Mg−1 digestate	Møller et al. [60]
Energy cost	USD 0.0689 kWh−1	Iowa industrial energy cost [63]
Other maintenance	10% of capital	Applied to all capital costs, except upgrading and injection

summarizes all operating cost calculations.

2.7. Renewable Fuel Identification Numbers and the Low-Carbon Fuel Standard

A source of income for biogas producers included in our model is RINs. RINs are bought and sold within fuel markets to meet compliance with the RFS [64]. Biogas producers that utilize manure qualify for D3 RINs, defined as cellulosic fuels. We used the average price of D3 RINs from 2016–2019 of USD 1.92 per gallon of ethanol equivalent [65]. This value was converted to a price per GJ using a gallon of ethanol’s energy content (0.08 GJ or 77,000 BTU), then this credit was applied to each scenario’s annual energy production.

California’s LCFS is a program designed to reduce the carbon footprint of the state’s transportation fuel supply through the use of low-carbon and renewable fuels [66]. This program allows for additional credits to be collected by the RNG producer. Utilizing spreadsheets titled “Biomethane from Anaerobic Digestion of Organic Waste” and “Biomethane from Anaerobic Digestion of Dairy and Swine Manure” provided by CARB, we calculated carbon intensity (CI) scores for compressed natural gas produced from the swine manure and corn stover inputs. Swine manure results in a negative carbon intensity equal to approximately −100 g CO_2eq MJ⁻¹. The CI score for corn stover was calculated to be approximately 14 g CO_2eq MJ⁻¹. These CI scores were then compared to the diesel compliance standard in 2030, set by CARB to be 80.36 CO_2eq MJ⁻¹. Corn stover saves 66 g CO_2eq MJ⁻¹ and swine manure saves 180 g. Total carbon emissions reduction values were calculated separately for the energy production from methane for each feedstock, then multiplied by the 2016–2019 average value of USD 136 Mg⁻¹ CO₂ emissions avoided [67] and summed for a total credit value. The value of LCFS credits ranged from USD 21 GJ⁻¹ in S1 to USD 16 GJ⁻¹ in S2-4. Both credits require the end use of the RNG is as a transportation fuel.

2.8. Final Cost Calculation

The minimum selling price (MSP) for each scenario was calculated by summing annualized capital and operating costs and subtracting the annualized RIN, LCFS, and digestate credit values. As mentioned in previous sections, costs were converted to 2019 USD. The year 2019 was chosen due to extreme price and inflation volatility since the beginning of the COVID-19 pandemic. Sensitivity analysis was completed by varying each of the model inputs individually by 1%. The change in final cost was observed, and the sensitivity coefficients (percent change in output divided percent change in input) were calculated for each input [68].

2.9. Natural Gas Comparison

The 2016–2019 Citygate price of natural gas in Iowa was used to compare to biogas cost, consistent with the timespan used to average LCFS and RIN credits. According to the US Energy Information Administration, this price is USD 3.91 per 1000 ft³ [9]. Using the average Iowa natural gas energy content of 39.8 MJ m⁻³ (1067 BTU ft⁻³; [69]), this translates to a cost of USD 3.47 GJ⁻¹. Figure 2 shows the average Citygate natural prices in Iowa since 1984. The World Bank [70] forecasts a 15% increase in natural gas commodity prices in the US in 2030, relative to 2019, and predicts it to continue to rise to a 31% increase in 2035 (constant USD (2010 = 100)). A sharp increase in price was seen after 2020 in Iowa, which may make the systems described in this paper more competitive than when compared to the 2016–2019 average.

3. Results and Discussion

3.1. Annual Production and Costs

Digester size, annual manure and corn stover inputs, annual digestate output, and the annual production of biogas, methane, and energy are summarized in

Table 5. Summary of digester size and annual inputs and outputs for the four modeled scenarios. S1 = single digester, stover amended; S2 = S1 w/water and stover amended; S3 = S2 w/centralized upgrading; S4 = centralized digester/upgrading, water and stover amended.

Parameter	S1	S2	S3	S4
Digester size (m3)	1.22 × 103	3.68 × 103	3.68 × 103	18.4 × 103
Manure processed (m3 yr−1)	7340	7340	7340	36,700
Corn stover processed (Mg yr−1)	375	1500	1500	7500
Solid digestate produced (Mg yr−1)	6.46 × 105	1.45 × 106	1.45 × 106	7.17 × 106
Biogas (m3 yr−1)	2.64 × 105	4.77 × 105	4.77 × 105	2.39 × 106
Methane (m3 yr−1)	1.53 × 105	2.78 × 105	2.78 × 105	1.39 × 106
Energy (GJ yr−1)	5780	10,500	10,500	52,300

. Because S1 utilizes less corn stover than the other scenarios, it is the lowest-production scenario. S2 and S3 have equal productions, and S4 has five times the production of S2 and S3, simply because it employs a single reactor receiving five times the feedstock.

Table 6. Overall and component costs of energy from each digester scenario. S1 = single digester, stover amended; S2 = S1 w/water and stover amended; S3 = S2 w/centralized upgrading; S4 = centralized digester/upgrading, water and stover amended.

Costs (USD GJ−1)	S1	S2	S3	S4
Manure	USD 7.66	USD 4.24	USD 4.24	USD 4.24
Corn stover	USD 3.31	USD 7.32	USD 7.32	USD 7.32
Manure transport	USD 0.00	USD 0.00	USD 0.00	USD 9.78
Capital	USD 53.91	USD 42.38	USD 53.63	USD 18.15
Operating costs	USD 10.33	USD 9.93	USD 12.13	USD 8.58
MSP (pre-credit)	USD 75.20	USD 63.87	USD 77.32	USD 48.08
Digestate credit	USD −7.87	USD −4.69	USD −4.69	USD −4.69
RIN	USD −23.63	USD −23.63	USD −23.63	USD −23.63
LCFS	USD −21.02	USD −15.79	USD −15.79	USD −15.79
MSP (including credits)	USD 22.68	USD 19.75	USD 33.20	USD 3.96

shows the MSP broken down by each cost category in the model (USD GJ⁻¹ basis). The higher volumetric carbon loading rate of S2 results in a 15% drop in the pre-credit raw cost per GJ energy produced compared to S1. In addition to the increased methane production from adding additional carbon through corn stover, the increased input mass and production required an increased size of digester and cleaning and injection equipment, leading to economy-of-scale advantages in S2 over S1. In contrast, the LCFS and digestate credits in S1 are greater than in S2 because less corn stover is added to the digester. Perhaps surprisingly, the centralized upgrading scenario (S3) was approximately 21% more expensive on a raw cost basis than S2. The economies of scale leveraged by centralized upgrading was overshadowed by the cost of transporting biogas to the centralized injection point via pipeline. S4 is the least costly scenario. The economy of scale associated with totally centralized production was more significant than the disadvantage of transporting manure, consistent with the results of Hengeveld et al. [26]. This result may vary with pipeline costs and biogas or manure transportation distances, which we explore further in Section 3.4

3.2. Capital

The unit capital costs of each component are summarized in

Table 7. Unit capital costs by component (USD GJ⁻¹) for each scenario. S1 = single digester, stover amended; S2 = S1 w/water and stover amended; S3 = S2 w/centralized upgrading; S4 = centralized digester/upgrading, water and stover amended.

Component	S1	S2	S3	S4
Digester	USD 14.00	USD 14.83	USD 14.83	USD 7.67
Upgrading and injection	USD 38.19	USD 26.60	USD 9.54	USD 9.54
Biogas transport	USD 0.00	USD 0.00	USD 28.31	USD 0.00
Storage and separation	USD 1.71	USD 0.95	USD 0.95	USD 0.95
Total	USD 53.91	USD 42.38	USD 53.63	USD 18.15

. The cost of upgrading and injection makes up the majority of the cost in three scenarios (S1, S2, and S4). Williams et al. [58] noted that total annual capital, operating, and maintenance costs for upgrading and injection can range from approximately USD 7 to 25 GJ⁻¹ when upgrading biogas for natural gas pipeline injection, depending on scale. Summing our capital and operating costs for upgrading and injection yields similar results, with unit costs from USD 12 GJ⁻¹ in S4 to 30 GJ⁻¹ in S1. Additionally, in S1 and S2, 500 m of pipeline to connect to the natural gas grid was included in the upgrading and injection cost category in Table 7. In S3, upgrading and injection cost is only 18% of the total capital cost due to the large investment cost for the biogas transportation pipeline, which accounts for over 50% of the total capital cost in S3. The portion of cost from the digester ranges from 25% in S1 to 42% in S4. Digester capital costs vary depending on digester type chosen and could be much lower if a different digester type (e.g., covered lagoon) is utilized. The strong economy of scale associated with digester capital cost resulted in unit digester costs in S4 being approximately half of the unit costs of the digester in S1, S2, or S3. The annual capital costs of S4 are significantly higher than in any other scenario, but those costs are more than offset by the higher energy production and economies of scale, making it the least expensive scenario on a per GJ basis.

3.3. Natural Gas Comparison

Figure 3 shows the scenario costs compared to the average price of natural gas in Iowa before (including the credit value of digestate as fertilizer) and after RIN and LCFS credits are applied. Before considering these significant credits, there are no scenarios that are competitive with the cost of natural gas. After applying the value of these credits, the costs become more competitive, with S4 being approximately USD 0.50 GJ⁻¹ more than the comparison natural gas price. The other three scenarios range from USD 16 to 30 GJ⁻¹ more than natural gas after applying RIN and LCFS values. These results highlight the heavy reliance on these credits for swine AD RNG systems to begin to become competitive with fossil fuels.

3.4. Effect of Transportation Distances

The cost impacts of transportation distance, either of biogas in S3 or of manure in S4, were explored by changing one-way distance over the range of 0 to 32 km (0–20 miles). The cost of the biogas pipeline was calculated three different ways. First, the cost was calculated using the USD 50,000 km⁻¹ (USD 1,000,000 mile⁻¹) capital cost utilized in the model. Then, the highest and lowest cost for the biogas pipeline according to Hengeveld et al. [26] were utilized instead (i.e., (USD 373,000 and USD 50,000 km⁻¹, respectively). As shown in Figure 4, the energy production costs of S3 exceeds that of S4 for all distances, unless the pipeline can be installed at 50,000 km⁻¹ (USD 80,000 mi⁻¹), which seems unrealistically low as this is 8% of the recommended cost estimation by the EPA [59]. At a manure transportation distance of 32 km (20 miles), the energy production cost in S4 increases by USD 54 GJ⁻¹ from the base case, greatly decreasing its competitiveness with the cost of natural gas. For centralized digestion systems similar to those modeled in this work to compete with fossil fuels, the transportation distance of the manure should be minimized. Swine farms in close proximity to each other with a central location for pipeline injection would be preferred, but this siting is not typical for other reasons [71]. For small-scale digestion with centralized biogas upgrading and injection, significantly lower pipeline costs are needed to be competitive with fossil fuels. Other methods of biogas transportation may prove to be a better choice.

3.5. Sensitivity Analysis

The sensitivity analysis results for the ten most sensitive model inputs in S4 are shown in Figure 5. For example, S4 had a base MSP of USD 3.96 (Figure 3); a 1% increase in biogas cleaning efficiency would lead to a 6.2% decrease in MSP (i.e., down to approximately USD 3.70). The 10 largest sensitivity coefficients ranged in magnitude from 3.5 to just under 13. Unsurprisingly, the highly non-linear digester scaling exponent, which drives the overall capital cost, had the highest sensitivity coefficient. The scaling exponent value was calculated by Faulhaber et al. [12] based on available AgSTAR data. While an increased scale generally decreases unit capital costs, the magnitude of this decrease depends on several different factors, such as location, unique design characteristics, and the type of digester selected. Additionally, the assumed efficiency of the digester was a highly sensitive input. This value was assumed to be 75% of the assumed methane potential of the inputs that could be reached. This value may increase or decrease based on environmental factors, such as weather.

The minimum selling price in S4 was very sensitive to both RIN and LCFS credit values, with a 1% change resulting in a 6.0% and 4.0% change in the final energy production cost, respectively. As shown in Figure 6, the value of both credits vary greatly over time and influence the competitiveness of S4 with natural gas. With no credits applied (red dotted line), S4 is never competitive with natural gas. With either credit applied independently, S4 is rarely competitive. Given the high sensitivity of the model results to these values, the volatility in their prices would profoundly affect the profitability of the system.

Less sensitive input assumptions in the top ten include the process assumptions related to input properties, namely methane production and VS content. These inputs may vary based on the age of the manure or specific livestock feeding information. The biogas cleaning efficiency was also among the most sensitive assumptions, which could change based on biogas cleaning methods deployed at the digester. The results were also sensitive to the annual operating days assumption, which was placed at 340 days yr⁻¹.

The impact of utilizing corn stover in the digesters was also explored. The benefit of adding corn stover varied with scale. When comparing S1 (farm-scale digester, no water added) to an identical scenario without corn stover additions, the price at which adding corn stover stops being advantageous is USD 70 Mg⁻¹. In a similar comparison, S4 (centralized digester with water addition) was compared to an identical scenario without corn stover (or water) addition. At the base price of USD 51 Mg⁻¹, adding corn stover decreases the MSP by approximately USD 4 GJ⁻¹. The price at which adding corn stover stops being advantageous in S4 is USD 81 Mg⁻¹. The wide-ranging prices for corn stover discussed above (USD 20–100 Mg⁻¹) could have a significant impact on the viability of the system. At a price of USD 100 Mg⁻¹, the MSP in S4 is approximately USD 7.50 GJ⁻¹ greater than the price of natural gas and is much less competitive.

4. Conclusions

This modelling effort suggests that centralized biogas production systems utilizing swine manure and corn stover are more cost effective than smaller-scale, decentralized digester systems. The diseconomies of scale associated with transporting manure and stover were smaller than the economy of scale associated with a larger digester and single cleanup and injection point. Critically, even the centralized scenario was highly reliant on transportation fuel-based greenhouse-gas-reduction credits to be competitive with fossil methane. This causes risk, because these credits are variable over time and frequently subject to policy changes. If on-farm (rather than centralized) digesters are used, a centralized biogas upgrading and injection point could only reduce the cost of production under very low biogas transport assumptions. Future research into low-cost biogas transport systems may be justified. Given the uncertainty of carbon credit prices, other incentives may be required to aid in the adoption of centralized systems. The environmental impacts of the system (harvesting corn stover, large volumes of manure in one location, methane leakage from the system) were not explored, but further research on these points is required to understand the overall costs of the system.