Conference Agenda

The production and utilization of fossil fuels have long been criticized for their wide-ranging impacts on various aspects such as environmental degradation. Conversely, developing clean energy technologies (CETs) is often hailed as a panacea to reduce dependence on fossil fuels and mitigate greenhouse gas emissions. Despite these perceived benefits, limited efforts have been made to evaluate the requirements and implications of the transition to CETs.

With the development of CETs, mining will emerge as a pivotal geostrategic sector. Therefore, estimating the material requirements for these technologies’ development is essential. Our research estimates the material requirements for the widespread adoption of CETs, assesses the capabilities to develop these technologies, and analyzes the recycling rates required to meet the projected capacity of CETs.

We start by translating the capacity of CETs forecasted by eight Integrated Assessment Models (IAMs) into the corresponding requirements for 36 key materials in the development of CETs. These include critical minerals, rare earth elements, platinum group materials, and structural materials. Our calculations reveal that meeting these projections requires scaling materials supply chains up at an unprecedented rate. When considering diverse technology types and their material requirements – information missing from IAMs – we find that this may represent a substantial 571-fold surge in selenium demand and a 531-fold increase in gallium, figures that seem difficult to achieve¹. This challenges the capacity of material reserves and the rate at which these can be produced. In turn, this diminishes the practical usefulness of IAMs, which are perceived as crucial tools in guiding academic discussions and shaping policy strategies.

To address this gap, we novelly adopt a bottom-up approach where we consider material availability constraints to estimate the capacity of CETs that could be realistically deployed. We find potential shortages compared to IAM projections that may result in deviations from the Paris Agreement target by 0.06–0.95 °C. At this point, IAM projections for CETs could still be met increasing material availability through recycling. However, we found that the recycling rate required for some materials like lithium would be above 300%, which does not seem an easy target to achieve.

Overall, this contribution quantifies potential shortages in technology capacities and the need to increase materials production rates. It also emphasizes the crucial role of incorporating these factors into IAMs for more accurate predictions and highlights the materials that developers should focus on. These findings provide crucial insights for evidence-based policymaking, aiming at a seamless transition towards sustainable energy systems.

References

1. Rostami F, Patrizio P, Jimenez Esteller L, Pozo Fernandez C, MacDowell N. Assessing the Realism of Clean Energy Projections. Energy Environ Sci. doi:10.1039/D4EE00747F

A key technology in integrating fluctuating, renewable energy in the process industry is the production of green hydrogen using water electrolysis plants. The scale-up of such electrolysis plant capacity remains a major challenge in achieving a successful transition to renewable energy. With this, the system simulation of these large-scale electrolysis plants can be utilized for process design but also for process monitoring and optimization, and maintenance scheduling (Mock et al., 2024).

Since the underlying processes for the simulation models are often not completely understood, hybrid modeling methods are a promising approach to combine process knowledge with process data for more reliable and precise simulation models (von Stosch et al., 2014). These hybrid, semi-parametric model achieved better accuracy than only knowledge-driven mechanistic models.

In this work a hybrid, semi-parametric model for an anion exchange membrane (AEM) electrolyzer module is developed. The basis of this model is a mechanistic model of the AEM stack (Große et al., 2024). The heat loss and heat transfer within stack, pump, heat exchanger and piping cannot directly be measured and therefore are approximated through the parameter estimation from real process data. Available sensors collect data for the temperature in the lye storage tank, the outlet temperature after the stack-unit, and the flow rate into the stack-unit. Using the available sensor data and the mechanistic stack model, the heat transfer coefficient and the heat loss within the peripheral components shall be estimated. The hybrid- semi-parametric model is then validated against real process plant data of the AEM stack module from different load settings of the electrolysis.

To evaluate the applicability of the hybrid, semi-parametric AEM model within a large-scale system simulation context, a large-scale electrolysis plant configuration is designed including multiple AEM stack modules, a water supply module and multiple post-processing steps for the produced hydrogen and oxygen. This plant configuration is then simulated using both the hybrid, semi-parametric and solely mechanistic AEM model to compare performance accuracy and simulation efficiency of both model types.

In future work, this hybrid, semi-parametric model could be the basis for creating more efficient system simulations utilizing surrogate models.

References

Große, M., Viedt, I., Lange, H., and Urbas, L., (2024). Electrolysis stack modeling for dynamic system simulation of modular hydrogen plants, International Journal of Hydrogen Energy. (in review)

Mock, M., Viedt, I., Lange, H., & Urbas, L. (2024). Heterogenous electrolysis plants as enabler of efficient and flexible Power-to-X value chains. In Computer Aided Chemical Engineering (Vol. 53, pp. 1885-1890). Elsevier.doi: 10.1016/B978-0-443-28824-1.50315-X.

von Stosch, M., Oliveira, R., Peres, J., and de Azevedo, S.F., (2014). Hybrid semi-parametric modeling in process systems engineering: Past, present and future, Computers & Chemical Engineering, 60, 86-101.

To meet the ambitious net zero target of the EU by 2050, it is a top priority to transition from fossil fuels to renewable sources. However, unlike traditional fossil fuel power plants, which can adjust output to match demand, non-dispatchable renewable sources like solar and wind are subject to natural variability and cannot be controlled to meet immediate demands. This is a challenge in the industrial sector where consistent and predictable energy usage is crucial. The EU project Flex4Fact aims to develop solutions to leverage energy and process flexibility in industry to meet a future with high renewable energy penetration.

A part of this project seeks to identify optimal investment strategies for enhancing energy flexibility, i.e. the capability of an industry to adapt to variable energy production. The investment strategies arise from energy system modelling, where the key is to understand how different technologies, such as solar power and batteries, complement the process flexibility. To enable these assessments, SINTEF’s open-source energy system model, EnergyModelsX [1], has been further developed to specifically address flexibility requirements at industrial sites. The considered flexibility aspects include process flexibility modelled as energy load shifting, allowing multiple energy carriers to cover the demand of single processes, and energy storage in terms of electric batteries. Moreover, synergies between these mentioned flexibilities and integrated on-site renewable expansion are focused upon within this work. Lastly, several sensitivity analyses are conducted to assess the robustness of the investment strategies towards changes in marked prices or scaled production.

The extended EnergyModelsX model is demonstrated through two case studies in the plastic and polymeric products manufacturing sector to evaluate their potential for increasing renewable generation and flexibility. The first use case, being energy intensive, consumes both natural gas and electricity. The main focus of this use case is heat recovery and utilization, hydrogen blending, on-site hydrogen production, and how this can reduce CO2 emissions. The second use case relies solely on electricity consumption, and the considered flexibility is energy shifting by electric batteries and production flexibility. The focus of this case study is on the interplay between energy storage, on-site energy production and process flexibility to increase the degree of self-produced renewable energy in the energy mix. Together, the two case studies demonstrate how the extended EnergyModelsX framework can be used to explore process and energy flexibility in industry to aid the transition from a fossil-based society to a renewable based society.

[1] L. Hellemo, E. F. Bødal, S. E. Holm, D. Pinel, and J. Straus, “EnergyModelsX: Flexible Energy Systems Modelling with Multiple Dispatch,” Journal of Open Source Software, vol. 9, no. 97, p. 6619, May 2024, doi: 10.21105/joss.06619.

Superstructure based approaches have long been a powerful method for optimal process synthesis problems. Specifically, there has been decades of research into the synthesis of distillation networks, due to their ubiquitous use in industry and the high cost of separation. The mathematical programs that result from such process synthesis problems are often large-scale mixed-integer nonlinear programs (MINLPs). These MINLPs are challenging to solve, even with state-of-the-art commercial solvers. Many solution approaches in literature rely on decomposition heuristics that exploit the structure of a specific problem formulation.

In this work we present a modeling approach for networks of thermally linked distillation columns. We address two of the major difficulties of previous superstructure approaches: problem generation and solution of the resulting mathematical program. For larger distillation network problems, generating the superstructure and index sets manually is cumbersome and error prone. We develop and employ an algorithmic approach for automatically generating state-task network superstructures and their corresponding index sets. The algorithmic approach allows for the generation of a problem of arbitrarily large size (N components in the system feed), limited only by computational cost.

Using the formulation for thermally coupled columns from Caballero & Grossmann (2004), the separation tasks and heat exchangers for a network of distillation columns are modeled using generalized disjunctive programming. Shortcut equations are used to model column behavior and aid in problem tractability. The FUG (Fenske-Underwood-Gilliand) equations can provide reasonable approximations of column behavior for multicomponent separations under certain assumptions. However, for large MINLPs, even models with Underwood equations can still be challenging to solve. After transformation to an MINLP, commercial solvers can fail to find solutions in reasonable computational times, even for small problem sizes. Through the introduction of intermediate variables, the Underwood equations can be reformulated from a general nonlinear form to bilinear equations. The resulting model is a mixed-integer quadratically constrained program (MIQCP). We can leverage recent advances in solver performance for quadratically constrained programs and solve the reformulated MIQCP using Gurobi where other solvers fail.

The ability to rapidly generate and solve large process synthesis problems is crucial in facilitating early-stage design decisions. Given the significant capital and operating costs associated with distillation networks, as well as their extensive application in the industry, optimizing these networks can lead to substantial energy and monetary savings. Our approach not only enhances the efficiency of solving such complex problems but also provides a scalable solution applicable to a wide range of process synthesis challenges.

References

Caballero, J. A., & Grossmann, I. E. (2004). Design of distillation sequences: From conventional to fully thermally coupled distillation systems. Computers & Chemical Engineering, 28(11), 2307–2329. https://doi.org/10.1016/j.compchemeng.2004.04.010

Sustainable industrial development depends on optimizing resource and energy integration within Eco-industrial parks (EIP), combined with stringent carbon emissions reduction policies. The main challenge is ensuring transparency, accountability, and data privacy while optimizing the conversion of raw materials and energy into valuable products and controlling emissions within EIPs. This research introduces an innovative framework to design optimized EIPs and deploy a blockchain-enabled trading platform for resources and emissions management, tackling these key issues. The proposed framework incorporates integrated EIPs combined with emission control policies, supported by two related systems: one for blockchain-based resource trading and the other for emissions control. The resource trading platform fosters transparency, enabling accurate tracking of material and energy flows. Furthermore, the framework integrates a Mixed-Integer linear Programming model (MINLP) with smart contracts on the Ethereum blockchain, ensuring data privacy, traceability, and equitable cost distribution among processes to meet environmental targets. The model also determines emission reductions and investments in carbon capture technology, promoting operational efficiency. Offering a powerful tool to decision-makers and authorities, this framework enhances comprehension of resource and emissions tracking, paving the way for the development of innovative policies and fostering regulatory compliance. This development underscores a leap in promoting sustainable industrial activities and aligning with environmental goals.

Natural gas processing is a crucial step in the Oil & Gas chain, gaining importance due to the high gas-oil ratio in the Pre-Salt layer and its role as a transition energy source. As midstream facilities, processing plants handle dynamic operations, a scenario intensified by Brazil's gas market opening, which allowed third-party access to sites like Petrobras’. This shift increased scheduling demands, requiring thousands of process simulations monthly and further enhancing their role. Short, medium, and long-term scheduling results drive corporate decisions across departments: Logistics, Market, Strategic Outlook, Performance Assessment, and Gas Planning & Optimization.

In this context, this work proposes an innovative digital-based scheduling tool for industrial application (IntegraGas: Integrated-Gas-Scheduling-Hub) that was tested as case study and is currently in use at Petrobras for: plant modeling, automation, integration and management of three distinct scheduling processes in four industrial gas processing assets. The framework implemented uses AspenHYSYS for process simulation, VBA to manage process simulator two-way communication, MicrosoftExcel as data transfer and end-user interface, PowerBI as analytics layer. IntegraGas includes several built-in features tailored for user experience and scheduling needs: 1) import/export data from/to third-party system, 2) select and automatically execute a sequence of what-if scenarios, 3) check legal limits and other product properties (automated warnings upon constraint violations), 4) freely configure plant setup and operation modes, 5) quick graph visualization and insights.

IntegraGas’s engineering core lies in first-principles process models, a virtual representation that was developed in the light each plant PFDs and validated against actual industrial data (average deviations 4-6%). The simulations were specifically tailored for scheduling application, ensuring appropriate compromise between model fit, execution time and open-market transparency requirements. Also, an extensive engineering work was carried out on mapping key process plant variables, i.e. the ones to which production scheduling was the most sensitive, and adding them as frontend user inputs (manipulated variables in the simulation models).

Historically, scheduling tools use overly-simplified models and manual flow of information in/out/within the process. This is mainly due to previous computation power limitations, that hindered the use of first-principles and plant-wide modeling, as well as the development of automated digital tools, which were also less needed in the aforetime slower-paced market. For current industry dynamics, though, this former path is not anymore satisfactory, as other needs arose particularly with the integration between scheduling tools and day-to-day strategic corporate decisions and online optimization tools such as process digital twins and process automation layers. This work comes as a breakthrough as it enters this new industry reality providing an integrated, robust and automated solution aligned with industry digitalization. The use of IntegraGas: 1) enabled fulfillment of scheduling processes for open-market contracts, avoiding company exposure to penalties; 2) provided efficiency gain, reducing in 24h the daily scheduling execution time (30 simulations in less than 1h) and providing the user appropriate time for critical output review; 3) increased reliability of engineering results and of data flow between company departments. Thus presenting itself as a groundbreaking tool towards a novel approach for gas processing scheduling.