The nth-plant scenario for blended feedstock conversion and preprocessing nationwide: Biorefineries and depots
Graphical abstract
Introduction
Feasibility studies of agricultural residue conversion to biofuels, bioproducts, and/or biopower are on the rise given biomass’ potential to become the major source of US renewable energy [1]. The goal is to mitigate the negative impact of climate change and provide energy security. Currently, the most widely produced biofuel is conventional ethanol (derived from corn starch) which is an effective substitute for fossil fuel in the transportation industry. The US is one of the largest fuel ethanol producers in the world with 200 plants that total a national name plate production capacity of over 16.9 billion gallons [2], 42% of the global biofuel production share [3]. To restrict competition of food resources and pressure on arable lands, US has limited the production of conventional biofuels to 15 billion gallons and set a target of 21 billion gallons per year (BGY) of non-edible feedstock to boost the total renewable fuel production by 2022, from which at least 16 BGY should be from cellulosic biofuels [4]. Unlike the food-based biomass resources, cellulosic biomass are non-edible resources including energy crops, municipal solid waste, and agricultural or forest residues [5]. Due to widespread availability and low-cost raw material, cellulosic biomass is a promising alternative for starch-based biomass. However, the cost of production of biofuels from cellulosic biomass is unclear due to the complex preprocessing operations, transportation and storage conditions [6].
Since the cellulosic biofuel production in the US was unable to meet the predictions for year to date, the Environmental Protection Agency (EPA) reduced the volume required to comply with RFS2 [4]. EPA had previously demanded 10.5 billion gallons of cellulosic biofuel production for year 2020, but had to reduce their targets to 590 million gallons [7]. This production shortage could be overcome with an efficient supply chain system. Currently, the cellulosic biofuel supply chain depends on the conventional/centralized supply system where feedstock is harvested, baled and stored locally close to the farms. Bales are then collected from farms and transferred directly to biorefineries. But, given the inherent characteristic of agricultural residues, such as non-flowable and bulky, this system is not efficient in handling cellulosic resources. Several studies have shown that, this system fails to handle supply regions with lower yield and larger supply area, which are often the case for cellulosic biomass [8], [9], [10]. They are complex to handle due to their dispersed geographic location, and quality variability. Therefore, the feedstock logistics for cellulosic biofuel constitutes 35–50% of the total production cost, which constraints the near-term development of a consistent market [11].
We posit that an advanced feedstock supply system that ensures the delivery of on-spec biomass at the gate of biorefinery would reduce production costs and ultimately accelerate the national biofuel industry [12], [13]. The idea is to move biomass-preprocessing operations from the biorefinery closer to the farmgate and into preprocessing depots. And, because these smaller facilities, when compared to biorefineries, could be built in the lower yielding regions not accessible by conventional biorefineries [14], depots will help increase the supply region of the supply chain system. These depots would receive biomass with heterogeneous characteristics and provenance from nearby supply regions for drying, grinding, and densification to a uniform format feedstock [12]. Pellets shipments to biorefineries would be based on a biomass blend/ratio with specified qualities including ash, moisture and carbohydrate content. Because the focus has been to design a cellulosic biofuel supply chain that maximizes quantities delivered at a biorefinery, very few studies have used the concept of biomass blending for on-spec deliveries [15], [16], [17], [18], [19]. Blending cost will depend on quality targets for different conversion pathways. Feedstock blend that meets target carbohydrate and ash content costs 12.12% higher than feedstock blend that meets only carbohydrate requirement [17].
When dealing with an advanced feedstock supply system, finding the location and size of the depots and biorefineries alongside with identifying the optimum feedstock blend and logistics cost, can ensure a long-term financial stability of the cellulosic biofuel production. Converting raw biomass into biofuel involves several stages including harvesting, baling, preprocessing, storage and transportation. A cost-competitive and efficient design of the biofuels supply chain requires the integration of the interdependencies and complexities of all the different stages. Numerous studies have been found in literature to model and optimize the cellulosic biofuel supply chain stages. Some of these studies considered a single feedstock [20], [21] while others considered multiple feedstocks [22], [23]. Almost all of the studies have considered a regional supply area instead of nationwide scenario. Gonzales et al. [24] developed a GIS-based heuristic to identify the depots and biorefineries throughout the US to locate the stranded and accessible herbaceous biomass. But, the study did not consider the on-spec delivery within target cost. Ekşioğlu et al. [25] used a mathematical model to identify the location, size and number of biorefineries as well as average travel distance and transportation costs to produce cellulosic ethanol from corn stover in Mississippi. Bai et al. [26] proposed Lagrangian Relaxation (LR) based heuristics to predict biorefinery locations in Illinois for optimum biorefinery investment, feedstock and transportation cost. Marvin et al. [23] developed a mixed-integer linear programming (MILP) model which can handle five different types of agricultural residues to determine the optimal location and size of biorefineries for a nine-state region in Midwestern US. Ng et al. [27] developed an MILP model with multi-year horizon to minimize total annual cost determining the optimal number, capacity and location of depots and biorefineries, the production inventory and shipment profiles. Corn stover and switchgrass was considered to use the model in Southern Wisconsin.
The feedstock cost can be divided into three groups, (1) grower payment, (2) logistics cost, and (3) quality costs. Most of the studies found in literature have tried to optimize the logistics cost while maximizing the supply. Delivering the optimal feedstock blend to the biorefinery considering both quality and quantity of feedstock, is still in its infancy in terms of research. Roni et al. [28] developed an MILP model to optimize feedstock sourcing decisions and depot locations while considering a least-cost blend formulation for multiple feedstock (agricultural residues, energy and municipal solid waste). The quality biomass parameters considered by Roni et al. were carbohydrate, ash and moisture content to identify the optimum feedstock blend to feed biorefinery in Kansas. Roni et al. only considered the supply chain for a single biorefinery while identifying the depot locations. Since cellulosic biomass is costly to handle and transport, higher production cost puts another limitation to the advancement of this industry alongside with the quality constraints. The Department of Energy’s (DOE) goal is to achieve a near-term $3/GGE by 2022 or a long-term goal of $2.5/GGE by 2030, where feedstock handling and delivery costs are $71.26/dry tons and $79.07/dry tons respectively at the biorefinery gate [29].
The authors identified a knowledge gap in the literature that no other studies have considered a nationwide delivery of on-spec biomass to biorefineries while meeting a target minimum fuel-selling price (MFSP) of $3/GGE by 2022 and $2.5/GGE by 2030. This study aims to fill in that research gap by optimizing both the logistics cost as well as the quality costs at the same time while handling the complexities of nationwide delivery under a target biofuel price. The novelty in this study is that we provide an economically and technically viable industry path to the development of a national biofuel industry by answering some of the key questions: (i) How much biomass can be delivered nationwide under the quality and cost target? (ii) What are the logistics cost required for delivery? (iii) What are the optimum locations and capacities of depots and biorefineries nationwide? And, (iv) What are the possible scenarios for various states in a depot-based system? To answer these questions, we developed a modified version of the least-cost formulation model [30] to deliver on-spec biomass while simultaneously optimizing the biorefinery and depot locations and nameplate capacities nationwide to meet the national MFSP target set by DOE. Contributions from this study can be summarized as followed:
- (i)
Validation that a larger supply radius and a higher quantity of biomass can be accessed using the advanced feedstock supply system with distributed depots to meet competitive biofuel prices.
- (ii)
Exemplary scenarios with a national mature conversion technology that takes advantage of economies of scale, the nth-plant scenario.
- (iii)
Finally, this study contributes to the literature with a nationwide database of field-depot and depot-biorefinery location and allocation considering multiple scenarios to meet DOE near- and long-term cost targets. The dataset will be available to other researchers for further analysis and decision-making purpose.
Section snippets
Model approach
The mixed-integer linear program (MILP) model presented was developed using the OPTMODEL procedure in SAS Institute Inc. 9.4 M4 and the branch and bound algorithm was used to solve the model. Fig. 1 represents the decision network used to formulate the advanced supply system and includes different farmgate price levels, feedstock types, field locations, depot locations, and biorefinery locations. The MILP analyzes the different biomass feedstock quantities available at various farmgate prices
Model formulation
The MILP model presented in this paper identifies the optimal location and size of an undetermined number of biorefineries and depots to maximize total feedstock (X) delivered to biorefineries at less than or equal to a specific target price (Eq. (1)). We analyzed two target prices: $79.07 and $71.26 per dry tons ($2016) based on the short- and long-term goals presented by a DOE techno-economic analysis [29]. Table 3 presents the data sets, parameters, and decision variables in our MILP
Scenarios
Four different scenario runs were performed considering the year and cost target, namely (S1) 2022 at $79.07/dry tons, (S2) 2030 at $79.07/dry tons, (S3) 2040 at $79.07/dry tons and (S4) 2030 at $71.26/dry tons. Even after decreasing the set of depot and biorefinery candidates, the problem had around 43,000 variables, 5,500 constraints and 16,000 constraint coefficients. We ran each scenario for 3 h and obtained an error gap between 0.17 and 17%. The results for the different years and targeted
Discussion
The main goal of the presented study was to analyze the nationwide scenario for cellulosic biofuel production and determine the feasibility of the EPA’s target of 16 billion gallons by year 2022. Considering a biofuel yield of 44.8 GGE/dt [29], around 357 million dt of feedstock needs to be delivered at the gate of the biorefinery and a total of 493 biorefineries with 725,000 dt capacity have to be built to meet EPA goals. However, the results of the developed model indicated that only 42.8
Conclusion
To provide economic sustainability for cellulosic crop production, the location of cellulosic based biomass depots and biorefineries have to be strategic throughout the US, creating sufficient cellulosic biomass demand in the market and reducing the pressure on food production. Findings from this study could be used to provide cost and profit analysis of cellulosic biofuel production to the decision-makers including supply managers, farmers and business investors which could ensure a
CRediT authorship contribution statement
Tasmin Hossain: Methodology, Software, Data curation, Writing - original draft, Visualization. Daniela Jones: Conceptualization, Methodology, Validation, Resources, Writing - original draft, Supervision, Project administration. Damon Hartley: Conceptualization, Methodology, Investigation, Resources, Writing - review & editing, Supervision, Data curation. L. Michael Griffel: Visualization, Conceptualization, Methodology, Data curation. Yingqian Lin: Conceptualization, Methodology, Validation.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This work was funded by the U.S. Department of Energy’s Office of Energy Efficiency and Renewable Energy (EERE) Bioenergy Technology Office under DOE Idaho Operations Office Contract DEAC07-05ID14517 with Battelle Energy Alliance, LLC, contract DE-AC05-00OR22725 with UT-Battelle, LLC, and USDA Hatch Project funds. The views expressed in this publication do not necessarily represent the views of the DOE or the U.S. Government. The U.S. Government retains and the publisher, by accepting the
Data availability
The input dataset for the supply curve of this study can be found using the County Download Tool at the Bioenergy Knowledge Discovery Framework open-source database [1]. The published dataset from the results of this study are attached with the manuscript.
References (35)
- et al.
Comparing alternative cellulosic biomass biorefining systems: Centralized versus distributed processing systems
Biomass Bioenergy
(2015) - et al.
Lignocellulosic biomass for bioethanol production: current perspectives, potential issues and future prospects
Prog Energy Combust Sci
(2012) - et al.
Techno-economic analysis of decentralized biomass processing depots
Bioresour Technol
(2015) - et al.
Simultaneous application of predictive model and least cost formulation can substantially benefit biorefineries outside Corn Belt in United States: A case study in Florida
Bioresour Technol
(2019) - et al.
Logistics system design for biomass-to-bioenergy industry with multiple types of feedstocks
Bioresour Technol
(2011) - et al.
GIS-based allocation of herbaceous biomass in biorefineries and depots
Biomass Bioenergy
(2017) - et al.
Analyzing the design and management of biomass-to-biorefinery supply chain
Comput Ind Eng
(2009) - et al.
Biofuel refinery location and supply chain planning under traffic congestion
Transp Res Part B: Methodol
(2011) - et al.
Design of biofuel supply chains with variable regional depot and biorefinery locations
Renewable Energy
(2017) - et al.
Distributed biomass supply chain cost optimization to evaluate multiple feedstocks for a biorefinery
Appl Energy
(2019)