Techno-economic assessment of biomass-based natural gas substitutes against the background of the EU 2018 renewable energy directive

In this work, membrane separation is chosen for biogas upgrading as it is compact, requires mainly electricity for compression, and yields CH₄ at high purity and pressure ready for injection. The feedstock mixture is chosen according to its availability and in such a way that competition with food production is minimized. The following two biomethane production plants were investigated:

Figure 1 shows a simplified flowchart of the biomethane production processes.

In order to create the design basis for the biogas plants, the feedstock mixture was used as input for the KTBL biogas plant calculator [17], which also provides detailed feedstock information, to calculate the biogas capacity of the digester. The feedstock is converted in the anaerobic digester into biogas. A partial flow of the biogas as well as of the retentate of the first stage of the membrane separation unit is used as fuel in a boiler to heat the digester to maintain the temperature at a sufficient level between 30 and 55 °C. Subsequently, the raw biogas is cooled, where steam is condensed and then fed into the pretreatment unit where minor components like H₂S and NH₃ are separated using activated carbon filters. The cleaned biogas is compressed to an absolute pressure of 10 bars and subsequently fed into the two-stage membrane separation unit. The retentate of the first membrane unit is fed back into the boiler as fuel and the retentate of the second membrane unit is recycled back to the compressor to increase the partial pressure of the CH₄. Finally, the upgraded biomethane is sent to the injection station for drying, lower heating value (LHV) adjustment, and odorization and then injected into the natural gas grid. The injection station is not included in the techno-economic assessment.

The investigated thermochemical process for BioSNG production is based on DFB steam gasification of wood chips [18, 19] and the VESTA SNG process from Amec Foster Wheeler [20]. Information regarding gas-cleaning equipment was taken from [11]. The investigated BioSNG capacities are 6.1, 12.2, and 49.1 MW. Figure 2 shows a simplified flowchart of the BioSNG process. The upper part shows the DFB gasification process including the product gas compression and the lower part shows the VESTA SNG process.

Mass and energy balances are based on the base case of the BioSNG plant (10 MW gasifier power) in [21] without considering district heat. Wood chips are fed into a dryer where the water content is reduced from 40 to 20%. Subsequently, the wood chips are fed into the gasifier, where they are converted with steam into product gas. The product gas, which is composed of about 40% H₂, 25% CO, 20% CO₂, 10% CH₄, and other components like higher hydrocarbons, sulfur components, and small amounts of N₂, is cooled, filtered, and fed into the RME (rapeseed methyl ester) scrubber where most of the tar, except BTEX (benzene, toluene, ethylbenzene, xylene), are removed and steam condensed. The condensate is reused for steam fluidization in the gasifier. After the RME scrubber, a partial flow of the product gas is extracted and used as fuel for the combustion reactor of the DFB system. The remaining product gas is led into regenerable fixed-bed adsorbers filled with activated carbon for removal of sulfur (especially H₂S) and BTEX. Then, the product gas is compressed to 10 bars and fed into the VESTA SNG process. In the first step, the product gas is led into the fixed-bed hydrogenation reactor where olefins are hydrogenated, and COS is converted into H₂S. The H₂S is subsequently removed by a ZnO adsorber bed. In the second step, the gas is fed with additional steam into a water gas shift (WGS) reactor where CO and H₂O react to H₂ and CO₂. Then the gas is cooled and led into the methanation section. Due to the highly exothermic methanation reaction, the section consists of several fixed-bed reactors in series. In each reactor, the CO and H₂ react to H₂O and CH₄. In addition, depending on operating conditions, CO and H₂O are formed from H₂ and CO₂ by the reverse WGS shift reaction. Subsequently, the gas enters an amine scrubber where CO₂ is removed. Finally, the gas enters the final methanation reactor where the remaining CO and H₂ are converted into H₂O and CH₄ at favorable operating conditions to reach the gas grid specifications before it is injected into the gas grid as BioSNG. Like the biomethane process, the injection station is not considered for the techno-economic assessment. The energy of the flue gas of the DFB system is recovered for steam generation which is needed for fluidization of the gasifier and for steam addition before the WGS reactor and combustion air preheating. The DFB gasification process operates at ambient pressure.

This section describes the methodology used for the techno-economic assessment of the investigated processes.

CAPEX were estimated based on a literature study as well as on budget quotes from different plant manufacturers. Investment costs from years other than 2016 were adjusted using the chemical engineering plant cost index (CEPCI; see Eq. 1). The assessment is carried out for the base year 2017 and it is assumed that 2016 CAPEX are equal to 2017 CAPEX.

First, the CAPEX of the smaller plant capacities were estimated. Subsequently, the CAPEX of the larger plant capacities were estimated using capacity rationing according to Eq. 2.

An exponent m of 0.67 was used as scaling factor for the BioSNG plants and biomethane plants as it is considered as good average [22]. The basis for the cost estimation was the inside battery limit (ISBL) costs, not taking the plant periphery, infrastructure, and site costs into account.

In addition, plant startup expenses (SUEX) were 10% of the calculated CAPEX. Therefore, the overall investment costs (INV) of a plant were calculated according to Eq. 3.

Tables 1 and 2 show the specific costs and prices of the different considered materials and energy streams as well as estimates for the calculation of the operating expenses (OPEX). These were taken from different sources after plausibility checks.

The annual depreciation was considered according to

Table 3 shows the assumptions for the techno-economic assessment covering the number of operators, their wage, the annual operating time, the overall plant lifetime, and the chosen return on investment (ROI).

It was assumed that the biomethane plants can be operated with one operator and the BioSNG plants need six operators to ensure a safe and reliable twenty-four-seven operation. In addition, a tax rate of 25% and a return on investment of 10% were chosen.

The following key figures were calculated to describe the techno-economics of the investigated processes.

The annual specific INV were calculated according to

The annual specific OPEX were calculated according to

The annual specific TOTEX are the sum of CAPEX_s (INV_s) and OPEX_s.

The before tax (BT) cash flow was calculated based on OPEX. It was calculated according to Eq. 8.

The after tax (AT) cash flow was calculated according to Eq. 9, which takes the BT cash flow, the tax rate, and depreciation into account.

The techno-economic assessment was based on the NPV, which was calculated with the AT cash flow, the discount rate (i or ROI), the plant lifetime, and the investment costs according to Eq. 3.

The natural gas substitute (BioSNG or biomethane) selling price at an NPV equal zero was calculated. In addition, sensitivity analyses with − 25 to + 25% of the annual operating time, the CAPEX, and the feedstock price were carried out.

The production efficiency of the investigated processes was calculated according to Eq. 11 considering the lower heating value (LHV) of the feedstock and the produced renewable methane.

The CO₂ equivalent emission calculation is based on summarizing the contributions of the major material and energy streams. The data are entirely based on available literature. To calculate the CO₂ equivalent emissions of the produced natural gas substitutes, the reference values in Table 4 were used.

Fehrenbach et al. [23] only take CO₂ emissions and a part of the CH₄ emissions into account. No losses from the digester are considered and only the CH₄ emissions from the digestate storage are considered (0.1% for closed storage and 2.7% for open storage, e.g., like in most small manure-based plants). Recent studies [24] showed that typical CH₄ emissions are higher and that the differences between single plants can be significant. Therefore, some of the GHG emissions in Table 4 might be too low. Other material and energy streams were neglected due to their low contribution. The fossil reference value is 83.8 g CO₂ equivalent per MJ natural gas substitute according to [16]. A 60% reduction according to the renewable energy directive 2018 leads to an allowed value of 33.5 g CO₂ equivalent per MJ natural gas substitute. With a higher value, the generated biomethane or BioSNG cannot be declared as biofuel.

Table 5 shows the mass flows of the investigated processes for biomethane and BioSNG production.

In general, the energetic efficiency (compare Eq. 11) of biomethane processes strongly depends on the employed feedstock. Based on the assumptions in this work, the biomethane processes reach energetic efficiencies of about 48 to 50% based on the ratio of the generated biomethane to the methane potential of the feedstock input [17]. The main sink of the energy is the digestate. Billig [25] gives conversion efficiencies based on the higher heating value between 39.8 and 73.8% depending on the used feedstock also employing membrane separation for the biogas upgrading.

The BioSNG processes reach a higher energetic efficiency (compare Eq. 11) of about 65% based on the ratio of the generated BioSNG to the wood chip input (wet). The energy losses are mainly thermal losses caused by the gasification process as well as of losses due to the heat of reaction in the water gas shift reactor and the methanation reactors. The results are in agreement with other published literature. Rehling et al. [13] simulated a DFB gasification-based SNG process using an isothermal SNG reactor and achieved a production efficiency of 66.0%. Furthermore, [26] simulated a combination of a DFB-based SNG process with addition of H₂ from a power to gas PEM electrolyzer where an overall efficiency of 64.2% was reached by considering the biomass input and the electricity input from the PEM electrolyzer. Moreover, at the GoBiGas plant, which employs the DFB steam gasification process and the fixed-bed methanation TREMP process from Haldor Topsoe, a SNG production efficiency of 61.8% is reached [11]. The lower value can be explained by the lower water content of the fuel (8% compared to 40% in this work) and the corresponding increase of the lower heating value, and the additional recycle of syngas to the SNG stages as well as the difference in CO₂ removal (upstream of the methanation reactors). Bunten [27] reports an SNG production efficiency of 62% converting municipal solid waste into SNG by employing the VESTA SNG process. Moreover, [28] reports a SNG conversion efficiency of woody biomass between 60 and 67% by using the VESTA SNG process.

Figure 3 shows the overall investment costs of the investigated plants including CAPEX and SUEX.

The absolute investment costs of the BioSNG plants are significantly higher than the investment costs of the biomethane plants. This can be dedicated to the more sophisticated process of the BioSNG plants with several reactors, heat exchangers, and the necessary gas-cleaning equipment. The estimation of the biomethane plants is based on [29, 30] and the estimation of the BioSNG plants is based on [21]. Both estimations include ISBL costs and the SNG compression to 10 bars. However, the grid injection station is not included.

Figure 4 shows the specific annual TOTEX of the investigated processes.

The specific TOTEX of the biomethane plants are lower than the specific TOTEX of the BioSNG plants. The specific investment costs and the CAPEX-related costs are always higher for the BioSNG plants. However, the highest specific raw material cost has the 4.8 MW biomethane process. In general, labor-related costs are insignificant for all plants. The figure also shows that CAPEX-related OPEX have the highest share in case of the smaller plant capacities (compare Table 2).

Figure 5 shows the BioSNG and biomethane selling prices in dependence of the plant capacity considering a plant lifetime of 20 years and a ROI of 10%. The areas indicate the respective sensitivity analyses. The sensitivity of the CAPEX was calculated between − 25 and + 25% and the sensitivity of the operating hours was calculated between the reference (8000 h a⁻¹) and − 25% (6000 h a⁻¹).

The selling price of the BioSNG, especially for the smaller plant capacities, is significantly higher than the selling price of biomethane. This can be dedicated to the significantly higher CAPEX and CAPEX-related OPEX. In addition, the utility costs are also significantly higher (compare Fig. 4).

Gassner and Maréchal [31] report SNG production costs based on biomass gasification from 76 to 107 EUR MWh⁻¹ for a fuel power of 20 MW and 59 to 97 EUR MWh⁻¹ for a fuel power of 150 MW in 2009. However, they assumed higher production efficiencies between 69 and 76%. Moreover, [25] reports BioSNG production costs of biomass steam gasification plants between 68 and 182 EUR MWh⁻¹ depending on plant capacity and feedstock for the reference year 2012. Biomethane from biogas upgrading plants has production costs from 57 to 129 EUR MWh⁻¹ depending on the feedstock and the upgrading unit in the reference year 2012 [25]. In comparison, the actual price for natural gas in Austria (January 2018) according to [32, 33] is about 42 EUR MWh⁻¹ (incl. taxes and other charges).

The results of the biomethane plants are in agreement with [25] who gives production costs between 66 and 127 EUR MWh⁻¹ for 8.0 MW respectively 1.7 MW biomethane output for different feedstock.

Figure 6 shows the results of the calculation of the GHG emissions.

The requirements of the 2018 renewable energy directive (RED) are met by the 1.0 MW biomethane plant and all BioSNG plant capacities. Electricity consumption (EU-mix) is the major contributor in all cases. Although most small manure-based plants have relatively high CH₄ emissions [24], typically a CH₄ credit is considered since less CH₄ emitted of the manure if it is used in a biogas plant instead of a simple storage. This leads to significantly lower GHG emissions than for grass as main feedstock. Therefore, the grass-based 4.8 MW biomethane plant does not meet the requirements of the directive. In general, the BioSNG plants show lower GHG emissions which can be dedicated to the higher production efficiency of the natural gas substitute and the used feedstock which causes significantly less CO₂ emissions (compare Table 4). In addition, in case of the BioSNG plants, it is feasible to omit CH₄ emissions, if state-of-the-art technologies like a central waste gas treatment and flares are applied. The installation of such technologies is legally mandatory and less cost intensive at larger plant capacities.

Co-feeding of residual materials with a negative price (for example, sewage sludge) would lower the feedstock costs of the investigated processes and hence increase the economic feasibility.

Figure 7 shows the sensitivity of the NG substitute selling price with respect to the feedstock price.

The 4.8 MW biomethane process has the highest sensitivity regarding a feedstock price variation. A feedstock price decrease of 100% would lower the NG substitute selling price by about 40%. In contrast, the 1.0 MW biomethane process has the lowest sensitivity regarding a feedstock price variation. The BioSNG processes are in between, showing the highest sensitivity for the 49.1 MW process. This can be dedicated to an increasing share of the variable costs with increasing plant capacity. These results are promising for utilization of cheap feedstock, e.g., biogenic waste materials or waste wood for the application in a thermochemical BioSNG production. The results in Fig. 7 also show the resistance of the process economics to feedstock price changes.

Figure 8 shows the influence of co-feeding of residual material respectively the influence of the feedstock price on the absolute NG substitute selling price.

If the feedstock costs could be reduced by 100%, the NG substitute selling price of the 49.1 MW BioSNG process would be lower than the reference price of the 4.8 MW biomethane plant. Nevertheless, as mentioned in Section 3.2, the natural gas price in Austria is 42 EUR MWh⁻¹ (incl. taxes and other charges), which is close to the selling price of the 4.8 MW biogas plant of 48 EUR MWh⁻¹ at − 100% feedstock costs.