Record Types
###### Records with Type: Preprint
Perspectives on the Integration between First-Principles and Data-Driven Modeling
William Bradley, Jinhyeun Kim, Zachary Kilwein, Logan Blakely, Michael Eydenberg, Jordan Jalvin, Carl Laird, Fani Boukouvala
November 7, 2021 (v1)
Keywords: gaussian process regression, hybrid modeling, Machine Learning, model calibration, neural networks, physics-informed machine learning
Efficiently embedding and/or integrating mechanistic information within data-driven models is essentially the only approach to simultaneously take advantage of both engineering principles and data-science. The opportunity for hybridization occurs in many scenarios, such as the development of a faster model of an accurate high-fidelity computer model; the correction of a mechanistic model that does not fully-capture the physical phenomena of the system; or the integration of a data-driven component approximating an unknown correlation within a mechanistic model. At the same time, different techniques have been proposed and applied in different literatures to achieve this hybridization, such as hybrid modeling, physics-informed Machine Learning (ML) and model calibration. In this paper we review the methods, challenges, applications and algorithms of these three research areas and discuss them in the context of the different hybridization scenarios. Moreover, we provide a comprehensive c... [more]
Optimization under uncertainty of a hybrid waste tire and natural gas feedstock flexible polygeneration system using a decomposition algorithm
Avinash Subramanian, Rohit Kannan, Flemming Holtorf, Thomas A. Adams II, Truls Gundersen, Paul I. Barton
November 1, 2021 (v1)
Keywords: Decomposition Algorithm, Optimization under uncertainty, Polygeneration system, Stochastic Programming, Waste Tire, Waste-to-Energy
Market uncertainties motivate the development of flexible polygeneration systems that are able to adjust operating conditions to favor production of the most profitable product portfolio. However, this operational flexibility comes at the cost of higher capital expenditure. A scenario-based two-stage stochastic nonconvex Mixed-Integer Nonlinear Programming (MINLP) approach lends itself naturally to optimizing these trade-offs. This work studies the optimal design and operation under uncertainty of a hybrid feedstock flexible polygeneration system producing electricity, methanol, dimethyl ether, olefins or liquefied (synthetic) natural gas. The recently developed GOSSIP software framework is used for modeling the optimization problem as well as its efficient solution using the Nonconvex Generalized Benders Decomposition (NGBD) algorithm. Two different cases are studied: The first uses estimates of the means and variances of the uncertain parameters from historical data, whereas the seco... [more]
At what pressure shall CO2 be transported by ship? An in-depth cost comparison of 7 and 15 barg shipping.
Simon Roussanaly, Han Deng, Geir Skaugen, Truls Gundersen
July 7, 2021 (v1)
Subject: Optimization
Keywords: Carbon Capture and Storage, CO2 shipping, CO2 transport, Optimal transport pressure, Technoeconomic Analysis
While pipeline transport traditionally has been regarded as the best option for CO2 transport due to its low cost over short distances and important economies of scale, interest in vessel-based transport of CO2 is growing. While virtually all recent literature has focused on low pressure transport (at 7 barg and -46°C), the issue of optimal transport conditions, in terms of pressure, temperature and gas composition, is becoming more relevant as carbon capture and storage chains based on ship transport move closer towards implementation.
This study focuses on an in-depth comparison of the two primary and relevant transport pressures, 7 and 15 barg, for annual volumes up to 20 MtCO2/y and transport distances up to 2000 km. We also address the impact of a number of key factors on optimal transport conditions, including (a) transport between harbours versus transport to an offshore site, (b) CO2 pressure prior to conditioning, (c) the presence of impurities and of purity constraints, and... [more]
A mathematical model for prediction of long-term degradation effects in solid oxide fuel cells
Mina Naeini, Haoxiang Lai, James S. Cotton, Thomas A. Adams II
June 15, 2021 (v1)
A mathematical model of long-term solid oxide fuel cell (SOFC) degradation is proposed, based on a cross-cutting meta-study of SOFC degradation research available in the open literature. This model is able to predict long-term SOFC performance under different operating conditions, and it accounts for the main degradation mechanisms, including: Ni coarsening and oxidation; anode pore size changes; degradation of anode and electrolyte conductivity; and sulfur poisoning. The results of the study indicate that SOFCs initially degrade quickly, but that the degradation rate diminishes significantly after approximately 1200 hours of operation. Consequently, the effects of different factors associated with degradation rate are investigated, including current density, temperature, and partial pressure of H2 in fuel source. Sensitivity analyses show that current density and H2 partial pressure have the highest and the lowest impact, respectively. In addition, the model has been developed to asse... [more]
Data-driven Spatial Branch-and-bound Algorithm for Box-constrained Simulation-based Optimization
Jianyuan Zhai, Fani Boukouvala
November 14, 2020 (v1)
Subject: Optimization
Keywords: Black-box Optimization, Branch-and-bound, Simulation-based Optimization
The ability to use complex computer simulations in quantitative analysis and decision-making is highly desired in science and engineering at the same rate as computation capabilities and first-principle knowledge advance. Due to the complexity of simulation models, direct embedding of equation-based optimization solvers may be impractical and data-driven optimization techniques are often needed. In this work, we present a novel data-driven spatial branch-and-bound algorithm for simulation-based optimization problems with box constraints, aiming for consistent globally convergent solutions. The main contribution of this paper is the introduction of the concept data-driven convex underestimators of data and surrogate functions, which are employed within a spatial branch-and-bound algorithm. The algorithm is showcased by an illustrative example and is then extensively studied via computational experiments on a large set of benchmark problems.
Techno-economic Assessment of Optimised Vacuum Swing Adsorption for Post-Combustion CO2 capture from Steam-Methane Reformer Flue Gas
Gokul Sai Subraveti, Simon Roussanaly, Rahul Anantharaman, Luca Riboldi, Arvind Rajendran
August 18, 2020 (v1)
Keywords: Carbon dioxide capture and storage, Metal Organic Framework, optimisation, Steam-methane reforming, Technoeconomic Analysis, vacuum swing adsorption
This study focuses on the techno-economic assessment integrated with detailed optimisation of a four step vacuum swing adsorption (VSA) process for post-combustion CO2 capture and storage (CCS) from steam-methane reformer dried flue gas containing 20 mol% CO2. The comprehensive techno-economic optimisation model developed herein takes into account VSA process model, peripheral component models, vacuum pump performance, scale-up, process scheduling and a thorough cost model. Three adsorbents, namely, Zeolite 13X and two metal-organic frameworks, UTSA-16 and IISERP MOF2 are optimised to minimise the CO2 capture cost. Monoethanolamine (MEA)-based absorption technology serves as a baseline case to assess and compare optimal techno-economic performances of VSA technology for three adsorbents. The results show that the four step VSA process with IISERP MOF2 outperforms other two adsorbents with a lowest CO2 capture cost (including flue gas pre-treatment) of 33.6 € per tonne of CO2 avoided an... [more]
On the application of shooting method for determining semicontinuous distillation limit cycles
August 17, 2020 (v1)
Keywords: Hybrid Dynamical System, Limit Cycle, Optimization, Process Design, Semicontinuous Distillation
Semicontinuous distillation is a new separation technology for distilling multicomponent mixtures.
This process was designed using design methodologies with heuristic components that evolved
over twenty years. However, the fundamental philosophy of these design methodologies, which
involves guessing, checking and then using a black-box optimization procedure to find the values
of the design variables to meet some performance criteria, has not changed. Mainly, to address the
problem of having a heuristic simulation termination criterion in the black-box optimization phase,
the single shooting method for semicontinuous distillation design was proposed in this study. We
envision that this is a first step in the transformation of the semicontinuous distillation design
process for obtaining optimal designs. We demonstrate the application of this method using two
case studies, which involve the separation of hexane, heptane and octane.
Comparison of Steel Manufacturing Off-Gas Utilization Methods via Life Cycle Analysis
July 31, 2020 (v1)
Subject: Other
Keywords: blast furnace gas, coke oven gas, combined cycle power plant, Life Cycle Analysis, methanol production
This study utilizes life cycle analysis to compare three steel manufacturing off-gas utilization systems: a status quo system, which produces electricity via a low-pressure steam turbine; a combined cycle power plant (CCPP) system, which produces electricity using gas and steam turbines; and a methanol (MeOH) system, which converts coke oven gas (COG) and blast furnace gas (BFG) into MeOH (CBMeOH). This research seeks to compare the environmental impacts of each system based on equivalent raw material inputs. Since the systems have different products, system expansion is used to ensure that they have the same outputs and are therefore comparable. The system boundary consists of a combination of cradle-to-gate and gate-to-gate boundaries. The environmental effects of each system are compared at five locations—Ontario, the USA, Finland, Mexico, and China—using TRACI, CML-IA baseline, ReCiPe2016, and IMPACT2002+ in SimaPro v9. The results show that in Ontario, Finland, and China, CBMeOH s... [more]
Supply Chain Monitoring Using Principal Component Analysis
Jing Wang, Christopher Swartz, Brandon Corbett, Kai Huang
July 16, 2020 (v1)
Keywords: monitoring, Multivariate Statistics, Supply Chain
Various types of risks exist in a supply chain, and disruptions could lead to economic loss or even breakdown of a supply chain without an effective mitigation strategy. The ability to detect disruptions early can help improve the resilience of the supply chain. In this paper, the application of principal component analysis (PCA) and dynamic PCA (DPCA) in fault detection and diagnosis of a supply chain system is investigated. In order to monitor the supply chain, data such as inventory levels, market demands and amount of products in transit are collected. PCA and DPCA are used to model the normal operating conditions (NOC). Two monitoring statistics, the Hotelling's T-squared and the squared prediction error (SPE), are used to detect abnormal operation of the supply chain. The confidence limits of these two statistics are estimated from the training data based on the $\chi^2$- distributions. The contribution plots are used to identify the variables with abnormal behavior when at le... [more]
Technoeconomic Analysis of a Waste Tire to Liquified Synthetic Natural Gas (SNG) energy system
Avinash S. R. Subramanian, Truls Gundersen, Thomas A. Adams II
June 1, 2020 (v1)
Keywords: CO2 capture,, Gasification, Rubber, Synthetic Natural Gas (SNG), Waste tire, Waste-to-Energy
Thermochemical conversion of solid wastes through gasification offers the
dual benefit of production of high-value fuels and environmentally friendly
waste disposal. In this paper, we propose a novel process for production of
liquified synthetic natural gas (SNG) from waste tires via a rotary kiln gasification process. We use a combination of experimental data available in the
open literature, first principles mathematical models and empirical models to
study three design cases (without CCS, with precombustion CCS and with
pre- and postcombustion CCS) in two locations (USA and Norway). The
thermodynamic, economic and environmental performance of the concept is
studied. The results show that minimum selling prices of 16.7, 17.5 and 19.9
$/GJ_LHV,SNG are required for USA and 20.9, 21.8 and 24.9$/GJ_LHV,SNG
for Norway. We note that these prices may become competitive under certain
regulatory conditions (such as recent public policy movement in British
Comprehensive Environmental Impact Assessment of a Combined Petroleum Coke and Natural Gas to Fischer-Tropsch Diesel Process
March 13, 2020 (v1)
Subject: Other
In this study, a well-to-wheels life cycle assessment was conducted to determine the environmental impacts from disposing of petroleum coke by converting it into liquid fuel. Specifically, three processes for converting petroleum coke and natural gas to Fischer Tropsch diesel were investigated, both with and without carbon capture and sequestration (CCS). Impact categories were calculated using the EPA’s TRACI 2.1 US-Canada 2008 midpoint method in SimaPro software. In addition, the impact of grid emissions on the overall process was assessed using two representative Canadian locations with high (Alberta) and low (Ontario) grid emissions. The results of each impact category were compared among the designs and against conventional petroleum and oil-sands derived diesel. Key findings showed that the proposed designs when operated using CCS in the low-emissions-grid location had lower life cycle GHG emissions than conventional petroleum and oil-sands derived diesel. Nevertheless, the vario... [more]
Techno-economic analysis of coke oven gas and blast furnace gas to methanol process with carbon dioxide capture and utilization
LINGYAN DENG DENG, Thomas Adams II
January 9, 2020 (v1)
Keywords: blast furnace gas, CO2 utilization and storage, COG desulphurization, Coke oven gas, Economic and sensitivity analysis, methanol production
This paper documents a process for converting coke oven gas (COG) and blast furnace gas (BFG) from steel refineries into methanol. Specifically, we propose the use of blast furnace gas (BFG) as an additional carbon source. The high CO2 and CO content of BFG make it a good carbon resource. In the proposed process, CO2 is recovered from the BFG and blended with H2O, H2, and CH4-rich COG to reform methane. Optimized amounts of H2O and CO2 are used to adjust the (H2 – CO2)/(CO + CO2) molar ratio in order to maximize the amount of methanol that is produced. In addition, the desulphurization process was modified to enable the removal of sulfur compounds, especially thiophene, from the COG. The process design and simulation results reported herein were then used to determine any potential environmental and economic benefits. This research is based on off-gas conditions provided by ArcelorMittal Dofasco, Hamilton, Ontario. In order to determine which conditions are most desirable for this retr... [more]
Synthesis of feasible heat exchanger networks using attainable regions
Avian Yuen
December 9, 2019 (v2)
Keywords: Attainable region, Energy recovery, Heat exchanger network synthesis, Heat integration, Process Synthesis
The attainable region (AR) is a region in a performance space in which all physically realizable reactor network designs must exist. ARs have been used since the 1960s for solving reactor network synthesis problems. The benefits of these methods are that the feasibility of a performance target can be assessed prior to running a synthesis routine, the solutions they give are guaranteed to be physically realizable, and a design can be made robust to uncertainties in feed and performance targets by assessing whether a solution and the range of its possible values lie within the AR, just to name a few. In this article, the theory of attainable regions is extended to bring these benefits to the heat exchanger network (HEN) synthesis problem. Basic properties of the HEN-AR are proven and a synthesis method using the AR is presented with examples.
Data Science-Enabled Molecular-to-Systems Engineering for Sustainable Water Treatment
Elvis Eugene, William Phillip, Alexander Dowling
October 11, 2019 (v3)
Keywords: Bayesian optimization, design of experiments, fit-for-purpose water, inverse materials design, materials informatics, superstructure optimization, uncertainty quantification
Growing social and economic pressures demand technological innovations that enable the widespread usage of unconventional sources of water. These challenges motivate the emerging fit-for-purpose paradigm, wherein water is provided at the precise quality level of the intended application. Unfortunately, to date, fundamental advances in materials and nanotechnology have been slow to advance this paradigm. Using examples from membrane science and engineering, we highlight the critical need to bridge research at the molecular and nano-scales with development at the device and systems-scales to fully realize sustainable fit-for-purpose water technology. Specifically, we present four opportunities for computing and data science to accelerate convergence of sustainable water research: materials informatics and inverse design, model-based design of experiments, superstructure optimization, and uncertainty quantification. As such, we highlight opportunities to collaboratively revolutionize mole... [more]
Techno-economic and environmental analyses of a novel, sustainable process for production of liquid fuels using helium heat transfer
September 26, 2019 (v2)
Keywords: Biomass, Carbonless heat, Dimethyl Ether, Fischer-Tropsch Synthesis, Gasification, Methane Reforming, Negative emissions
In this paper, several new processes are proposed which co-generate electricity and liquid fuels (such as diesel, gasoline, or dimethyl ether) from biomass, natural gas and heat from a high temperature gas-cooled reactor. This carbonless heat provides the required energy to drive an endothermic steam methane reforming process, which yields H2-rich syngas (H2/CO > 6) with lower greenhouse gas emissions than traditional steam methane reforming processes. Since downstream Fischer-Tropsch, methanol, or dimethyl ether synthesis processes require an H2/CO ratio of around 2, biomass gasification is integrated into the process. Biomass-derived syngas is sufficiently H2-lean such that blending it with the steam methane reforming derived syngas yields a syngas of the appropriate H2/CO ratio of around 2. In a prior work, we also demonstrated that integrating carbonless heat with combined steam and CO2 reforming of methane is a promising option to produce a syngas with proper H2/CO ratio for Fisch... [more]
Side Stream Control in Semicontinuous Distillation
October 18, 2018 (v1)
Keywords: dynamic optimization, dynamical system analysis, Semicontinuous Distillation, side stream control
The idea to reduce cycle time (𝑇), by controlling the side stream flow rate using a feedforward control model -- the ideal side draw recovery arrangement (ISR) -- was standard in most semicontinuous distillation studies. However, its effect, particularly on ‘𝑇’ and more broadly on the system dynamics, was not clearly understood. In the current study, we compare the performance of using a modified form of ISR model with the status quo, based on the criteria, 𝑇 and separating cost (SC) on different case studies. Results show that the modified control model performed better with a 10-20% reduction in SC while maintaining product purities. Furthermore, the side stream flow rate trajectory that minimizes SC was found by using dynamic optimization and it did not differ a lot from the trajectory generated by the modified control model. The improvement in SC was at most 2%.
Optimization of Coke Oven Gas Desulphurization and Combined Cycle Power Plant Electricity Generation

LINGYAN DENG, Thomas A. Adams II
September 12, 2018 (v3)
Subject: Optimization
Keywords: carbon tax, coke oven gas valorization, combined cycle power plant, desulphurization, net present value, Optimization, steel refinery
Many steel refineries generate significant quantities of coke oven gas (COG), which is in some cases used only to generate low pressure steam and small amounts of electric power. In order to improve energy efficiency and reduce net greenhouse gas emissions, a combined cycle power plant (CCPP) where COG is used as fuel is proposed. However, desulphurization is necessary before the COG can be used as a fuel input for CCPP. Using a local steel refinery as a case study, a proposed desulphurization process is designed to limit the H2S content in COG to less than 1 ppmv, and simulated using ProMax. In addition, the proposed CCPP plant is simulated in Aspen Plus and is optimized using GAMS to global optimality with net present value as the objective function. Furthermore, carbon tax is considered in this study. The optimized CCPP plant was observed to generate more than twice the electrical efficiency when compared to the status quo for the existing steel refinery. Thus, by generating more e... [more]

Combining Petroleum Coke and Natural Gas for Efficient Liquid Fuels Production
Ikenna J Okeke, Thomas A Adams II
August 28, 2018 (v1)
This work explores the technical feasibility and economic profitability of converting petroleum coke (petcoke) and natural gas to liquid fuels via Fischer-Tropsch synthesis. Different petcoke conversion strategies were examined to determine the conversion pathway which can be competitive with current market prices with little or no adverse environmental impacts. Three main design approaches were considered: petcoke gasification only, combined petcoke gasification and natural gas reforming through traditional processing steps, and combined petcoke gasification and natural gas reforming by directly integrating the gasifier’s radiant cooler with the gas reformer. The designs investigated included scenarios with and without carbon capture and sequestration, and with and without CO2 emission tax penalties. The performance metrics considered included net present value, life cycle greenhouse gas emissions, and the cost of CO2 avoided. The design configuration that integrated natural gas refor... [more]
Biomass-Gas-and-Nuclear-To-Liquids (BGNTL) Processes Part I: Model Development and Simulation
James Alexander Scott, Thomas Alan Adams II
August 7, 2018 (v1)
New polygeneration processes for the co-production of liquid fuels (Fischer-Tropsch liquids, methanol, and dimethyl ether) and electricity are presented. The processes use a combination of biomass, natural gas, and nuclear energy as primary energy feeds. Chemical process models were created and used to simulate candidate versions of the process, using combinations of models ranging from complex multi- scale models to standard process flowsheet models. The simulation results are presented for an Ontario, Canada case study to obtain key metrics such as efficiency and product conversions. Sample Aspen Plus files are provided in the supplementary material to be used by others.
Deterministic Global Optimization with Artificial Neural Networks Embedded
Global deterministische Optimierung von Optimierungsproblemen mit künstlichen neuronalen Netzwerken
Artur M Schweidtmann, Alexander Mitsos
October 15, 2018 (v2)
Subject: Optimization
Artificial neural networks (ANNs) are used in various applications for data-driven black-box modeling and subsequent optimization. Herein, we present an efficient method for deterministic global optimization of ANN embedded optimization problems. The proposed method is based on relaxations of algorithms using McCormick relaxations in a reduced-space [\textit{SIOPT}, 20 (2009), pp. 573-601] including the convex and concave envelopes of the nonlinear activation function of ANNs. The optimization problem is solved using our in-house global deterministic solver MAiNGO. The performance of the proposed method is shown in four optimization examples: an illustrative function, a fermentation process, a compressor plant and a chemical process optimization. The results show that computational solution time is favorable compared to the global general-purpose optimization solver BARON.
Technical challenges in operating an SOFC in fuel flexible gas turbine hybrid systems: Coupling effects of cathode air mass flow
Nor Farida Harun, David Tucker, Thomas A. Adams II
June 19, 2018 (v1)
Keywords: Cathode air mass flow, Cyber-physical simulations, Fuel cell gas turbine hybrid, Fuel composition changes, Open loop characterization, Solid Oxide Fuel Cells
Considering the limited turndown potential of gasification technologies, supplementing a fuel cell turbine hybrid power system with natural gas provides flexibility that could improve economic viability. The dynamic characterization of fuel composition transients is an essential first step in completing the system identification required for controls development. In this work, both open loop and closed loop transient responses of the fuel cell in a solid oxide fuel cell (SOFC) gas turbine (GT) hybrid system to fuel composition changes were experimentally investigated using a cyber-physical fuel cell system. A transition from methane lean syngas to methane rich gases with no turbine speed control was studied. The distributed performance of the fuel cell was analyzed in detail with temporal and spatial resolution across the cell.

Dramatic changes in fuel cell system post combustor thermal output or “thermal effluent” resulting from anode composition changes drove turbine transients th... [more]
Space-constrained purification of dimethyl ether through process intensification using semicontinuous dividing wall columns
Sarah E. Ballinger, Thomas A. Adams II
June 12, 2018 (v1)
Keywords: Aspen Plus, Dimethyl Ether, Dividing wall column, Mobile Plant, Plant-on-a-truck, Process Intensification, Semicontinuous Distillation, Simulation
In this work, a distillation system is designed to purify dimethyl ether (DME) from its reaction by-products in the conversion of flare gas into a useful energy product. The distillation equipment has a size constraint for easy transportation, making process intensification the best strategy to efficiently separate the mixture. The process intensification distillation techniques explored include the dividing wall column (DWC) and a novel semicontinuous dividing wall column (S-DWC). The DWC and the S-DWC both purify DME to fuel grade purity along with producing high purity waste streams. An economic comparison is made between the two systems. The DWC is a cheaper method of producing DME however the purity of methanol, a reaction intermediate, is not as high as the S-DWC. Overall, this research shows that it is possible to purify DME and its reaction by-products in a 40-foot distillation column at a cost that is competitive with Diesel.
Modeling and simulation of an integrated steam reforming and nuclear heat system