FG Prozess- und Anlagentechnik
Refine
Document Type
Way of publication
- Open Access (13)
Keywords
- CO2 methanation (8)
- Biogas upgrading (3)
- Biomethane production (3)
- Waste valorization (3)
- CO2 waste valorization (2)
- CeO2 (2)
- Cu catalysts (2)
- Green energy (2)
- Hydrothermal carbonization (2)
- Mixed-integer linear programming (2)
Institute
BTU
In this work, the influence of uncertain parameters on the maintenance scheduling of Reverse Osmosis Networks (RONs) is explored. Based on a foundation of successful applications in various maintenance optimization domains, this paper extends the methodology to the domain of RON regeneration actions planning, highlighting its adaptability to diverse areas of dynamic processes with planning uncertainty. Traditional approaches in membrane cleaning scheduling have predominantly relied on MixedInteger Nonlinear Programming (MINLP), often leading to combinatorial problems that fail to capture the dynamic nature of the system. As part of this study, a novel approach based on the Multistage Integer Nonlinear Optimal Control Problem (MSINOCP) formulation is used to automate and optimize membrane cleaning scheduling without requiring combinatorial optimization. To evaluate the consequences of parameter uncertainty, 26 scenarios are considered in which the cost of the energy unit is considered as variable based on a random distribution, and these results are compared to a scenario where a fixed cost parameter is assumed. The findings show that when the cost of energy is considered as an uncertain parameter, the optimization process requires more frequent cleaning measures.
Control loop design, as well as controller tuning, constitute the pillars of process control to achieve design specifications and smooth process operation, and to meet predefined performance criteria. Currently, state-of-the-art approaches have focused on methods that yield only the pairings between input and output methods, and are not able to incorporate path and end-point constraints. This work introduces a novel strategy based on the multistage optimal control formulation of the control loop selection problem. This approach overcomes the drawbacks of traditional methods by producing an automated integrated solution for the task of control loop design. Furthermore, it obviates the need for any form of combinatorial optimization and incorporating path and terminal constraints. The results show that the proposed solution framework produces the same control loops as in the case of traditional approaches, however the inclusion of path and end-point constraints improves the performance of the control profiles.
Engineering the interfacial interaction between the active metal element and support material is a promising strategy for improving the performance of catalysts toward CO2 methanation. Herein, the Ni-doped rare-earth metal-based A-site substituted perovskite-type oxide catalysts (Ni/AMnO3; A = Sm, La, Nd, Ce, Pr) were synthesized by auto-combustion method, thoroughly characterized, and evaluated for CO2 methanation reaction. The XRD analysis confirmed the perovskite structure and the formation of nano-size particles with crystallite sizes ranging from 18 to 47 nm. The Ni/CeMnO3 catalyst exhibited a higher CO2 conversion rate of 6.6 × 10−5 molCO2 gcat−1 s−1 and high selectivity towards CH4 formation due to the surface composition of the active sites and capability to activate CO2 molecules under redox property adopted associative and dissociative mechanisms. The higher activity of the catalyst could be attributed to the strong metal–support interface, available active sites, surface basicity, and higher surface area. XRD analysis of spent catalysts showed enlarged crystallite size, indicating particle aggregation during the reaction; nevertheless, the cerium-containing catalyst displayed the least increase, demonstrating resilience, structural stability, and potential for CO2 methanation reaction.
Deep neural networks (DNNs) are frequently employed for information extraction in big data applications across various domains; however, their application in real-time industrial systems is hindered by constraints such as limited computational, storage capacity, energy availability, and time constraints. This contribution introduces the development of a novel hierarchical multiscale framework for the training of DNNs that incorporates neural sensitivity analysis for the automatic and selective training of neurons evaluated to be the most effective. This alternative training methodology generates local minima that closely match or surpass those achieved by traditional approaches, such as the backpropagation method, utilizing identical starting points for comparative purposes.
Catalytic Fast Pyrolysis of Lignocellulosic Biomass: Recent Advances and Comprehensive Overview
(2024)
Using biomass as a renewable resource to produce biofuels and high-value chemicals through fast pyrolysis offers significant application value and wide market possibilities, especially in light of the current energy and environmental constraints. Bio-oil from fast-pyrolysis has various conveniences over raw biomass, including simpler transportation and storage and a higher energy density. The catalytic fast pyrolysis (CFP) is a complex technology which is affected by several parameters, mainly the biomass type, composition, and the interaction between components, process operation, catalysts, reactor types, and production scale or pre-treatment techniques. Nevertheless, due to its complicated makeup, high water and oxygen presence, low heating value, unstable nature, elevated viscosity, corrosiveness, and insolubility within conventional fuels, crude bio-oil has drawbacks. In this context, catalysts are added to reactor to decrease activation energy, substitute the output composition, and create valuable compounds and higher-grade fuels. The study aim is to explore the suitability of lignocellulosic biomasses as an alternative feedstock in CFP for the optimization of bio-oil production. Furthermore, we provide an up-to-date review of the challenges in bio-oil production from CFP, including the factors and parameters that affect its production and the effect of used catalysis on its quality and yield. In addition, this work describes the advanced upgrading methods and applications used for products from CFP, the modeling and simulation of the CFP process, and the application of life cycle assessment. The complicated fluid dynamics and heat transfer mechanisms that take place during the pyrolysis process have been better understood due to the use of CFD modeling in studies on biomass fast pyrolysis. Zeolites have been reported for their superior performance in bio-oil upgrading. Indeed, Zeolites as catalyses have demonstrated significant catalytic effects in boosting dehydration and cracking process, resulting in the production of final liquid products with elevated H/C ratios and small C/O ratios. Combining ex-situ and in-situ catalytic pyrolysis can leverage the benefits of both approaches. Recent studies recommend more and more the development of pyrolysis-based bio-refinery processes where these approaches are combined in an optimal way, considering sustainable and circular approaches.
The implementation of novel CO2 valorization technologies is one of the most promising approaches towards the achievement of sustainable energy models. This chapter highlights the importance of carbon capture and utilization technologies and proposes novel approaches for the valorization of CO2-rich feedstock derived from thermochemical biomass conversion through the production of syngas mixtures via the Reverse Water Gas Shift reaction. After, this classification of the different types of nonconventional gases and biomass-treatment processes, we have also revised the fundamentals of the Reverse Water Gas Shift reaction and the impact of species commonly present in CO2-rich streams on the performance of the catalytic systems are also reviewed. Finally, a catalytic bi-functionalization approach that ensures larger CO productivity from simulated biomass-derived CO2-rich feedstock is demonstrated.
The reuse potential for the large annual production of spent coffee grounds (SCGs) is underexploited in most world regions. Hydrochars from SCGs produced via hydrothermal carbonization (HTC) have been recognized as a promising solid fuel alternative. To increase demand, optimization of the HTC and two post-treatment processes, washing and agglomeration, were studied to improve hydrochar in terms of energetic properties, minimizing unwanted substances, and better handling. HTC experiments at three scales (1–18.75 L) and varying process conditions (temperature T (160–250 °C), reaction time t (1–5 h), and solid content %So (6–20%) showed that the higher heating value (HHV) can be improved by up to 46%, and most potential emissions of trace elements from combustion reduced (up to 90%). The HTC outputs (solid yield—SY, HHV, energy yield—EY) were modeled and compared to published genetic programming (GP) models. Both model types predicted the three outputs with low error (<15%) and can be used for process optimization. The efficiency of water washing depended on the HTC process temperature and type of aromatics produced. The furanic compounds were removed (69–100%; 160 °C), while only 34% of the phenolic compounds (240 °C) were washed out. Agglomeration of both wet SCG and its hydrochar is feasible; however, the finer particles of washed hydrochar (240 °C) resulted in larger-sized spherical pellets (85% > 2000–4000 µm) compared to SCGs (only 4%).
Among challenges implicit in the transition to the post–fossil fuel energetic model, the finite amount of resources available for the technological implementation of CO2 revalorizing processes arises as a central issue. The development of fully renewable catalytic systems with easier metal recovery strategies would promote the viability and sustainability of synthetic natural gas production circular routes. Taking Ni and NiFe catalysts supported over γ-Al2O3 oxide as reference materials, this work evaluates the potentiality of Ni and NiFe supported biochar catalysts for CO2 methanation. The development of competitive biochar catalysts was found dependent on the creation of basic sites on the catalyst surface. Displaying lower Turn Over Frequencies than Ni/Al catalyst, the absence of basic sites achieved over Ni/C catalyst was related to the depleted catalyst performances. For NiFe catalysts, analogous Ni5Fe1 alloys were constituted over both alumina and biochar supports. The highest specific activity of the catalyst series, exhibited by the NiFe/C catalyst, was related to the development of surface basic sites along with weaker NiFe–C interactions, which resulted in increased Ni0:NiO surface populations under reaction conditions. In summary, the present work establishes biochar supports as a competitive material to consider within the future low-carbon energetic panorama.
Synthesis and Characterizations of Ni-doped Perovskite-Type Oxides for Effective CO2 methanation
(2023)
This work proposes Ni metal supported over rare earth-based emerging perovskite-type oxides as potential catalysts for the CO2 methanation. Presence of oxygen vacancies in perovskite-like materials enable them to exhibit higher catalytic activity. Furthermore, to tune the surface basicity, metal-support interaction and to enhance the activation of CO2, rare earth metals (La, Ce, etc.) are considered best candidates. Moreover, different perovskite-type supports (AxMnxO3, A= La, Ce) based on A-side substitution of rare earth metals were prepared with Ni metal loading of 10 wt.% via impregnation method.
Classical computing has experienced rapid growth in computational power, driven by the need to address increasingly complex industrial problems. The domain of global optimization plays a vital role in various applications, including optimal control, scheduling and assignment problems, or machine learning parameter selection. Currently, deterministic optimization techniques based on classical computing fail to deliver reasonable solutions within practical time constraints. Consequently, reliance on heuristic methods becomes common, albeit with no guarantee of solution quality. While ongoing algorithmic refinements lead to gradual enhancements in global optimization, they do little to address the fundamental issue of computational intractability. With the advent of quantum computing, a natural question arises: Can quantum methods offer advancements beyond classical approaches? Quantum annealing emerges as a promising subfield within quantum computing, necessitating the reformulation of problems as quadratic unconstrained binary optimization (QUBO) problems. In this contribution, a novel approach is introduced to transform relevant problems in Chemical Engineering into QUBO at two distinct levels of granularity. Subsequently, these problem systems are embedded within virtual quantum machines employing two different architectures. Additionally, a comparative analysis is performed, wherein the same problem is solved utilizing both classical global optimization methods based on metaheuristics and a hypothetical quantum annealer. The findings indicate that annealing-based solving methods exhibit the most potential, indicating their applicability to the transformed formulation Chemical Engineering problems.
In this work, cosolvent-stabilized superhydrophobic, highly hydrostable ZIF-67 was synthesized at room temperature using a facile, one-pot hydrothermal synthesis route, and the effect of cosolvent concentration on ZIF-67 crystal structure properties and hydrostability was studied systematically. The underlying mechanism for the cosolvent-supported hydrostability improvement was also proposed. Furthermore, the influence of hydrotreatment on the resultant ZIF-67s' catalytic performance was studied in the ‘Sabatier reaction’ for CO2 to synthetic natural gas (CH4) conversion.
Automotive technology is increasingly determined by drives based on electric motors in combination with batteries. The lithium-ion traction battery is a storage medium that combines high electrical efficiency with compact dimensions and relatively low weight. For the recycling of the cathode coatings (esp. Ni, Mn, Co) and peripheral battery components a variety of recycling options already exist. The graphite coating of the anodes has hardly been the focus of research activities to date. State of the art is currently the melting of the complete Copper-anode foils including graphite coating, whereby the graphite contributes only as a carbon carrier to the recycling of the copper. Separation and reuse of the very high-quality graphite on an industrial scale has not yet taken place.
At the BTU, a methodology has been developed, with which recovered anode graphites from traction batteries can be comprehensively characterised chemically and mechanically-physically. On this basis, targeted preparation for secondary applications is possible. The secondary graphites achieve a quality
that allows them to be reused as second-use anode material and for other applications.
A sustainable way to generate hydrogen is through dry biogas reforming, which uses methane gas and carbon dioxide to produce hydrogen. This study reveals partial results of the dry reforming of biogas in NiO-MxOy-Al2O3 catalysts (M=Na, K, Ca and Mg). The CO2 conversion varied between 79% and 94%, the CH4 conversion between 58% and 75%, the H2/CO ratio between 0.98 and 1.15 and the H2 yield between 37% and 45%. These values surpass literary references and the industrial catalyst, highlighting the promise of these materials for sustainable hydrogen production. The catalyst with Ca stood out due to its higher surface basicity, exhibiting the best conversion results and yield in H2.
different perovskite-type supports considering ABO3 (such as A= Al, La with B=Ce and A=Mg, Mn with B=Zr) were prepared via the sol-gel method. Ni metal loading of 10 wt.% was deposited on prepared perovskite supports via the impregnation method. The catalysts were characterized using XRD and FTIR techniques. The DRM activity was carried out in a tubular reactor as described in our previous study [5]. The catalytic performance was assessed in the temperature range of 500–700 ◦C, CH4/CO2 = 1/1 and under GHSV of 12,000 h–1. Among the prepared catalysts, Ni-doped perovskite combination (i.e. A=Mg with B=Zr)O3-δ exhibited higher (CH4, CO2) conversion ca. (69, 59) percent and syngas yield of ca. (H2/CO =0.72) at 700 oC. This indicates that the magnesium zirconate perovskite catalyst established strong interfacial metal-support interaction, redox properties and surface basic sites that linked with good performance of the catalyst during DRM process.
The urgent need to transform the energy landscape to achieve a sustainable future is underscored by policy commitments aimed at establishing a net-zero carbon economy by the year 2050. Within the EU, primary emphasis is directed towards the abatement of over 75% of emissions arising from electricity production, heat generation, transport, and industrial processes. In this work, technologies will be presented to engineer a low-carbon future.
Traditional food supply chains are often centralised and global in nature, entailing substantial resource consumption. However, in the face of growing demand for sustainability, this strategy faces significant challenges. Adoption of localised supply chains is deemed a more sustainable option, yet its efficacy requires verification. Supply chain analytics methodologies provide invaluable tools to guide decisions regarding inventory management, demand forecasting and distribution optimisation. These solutions not only enhance facilitate operational efficiency, but also pave the way for cost reduction, further aligning with sustainability objectives. This research introduces a novel decision-making approach anchored in mixed integer linear programming (MILP) and neighbourhood flow models defined in cellular automata to compare the environmental benefits and vulnerability to disruption of these two chain configurations. Additionally, a comprehensive cost analysis is integrated to assess the economic feasibility of incorporating layout changes that enhance supply chain sustainability. The proposed framework is applied on an ice cream supply chain across England over a one-year timeframe. The findings indicate the superiority of the localised configuration in terms of economic benefits, leading to savings exceeding £ 1 million, alongside important reductions in environmental impact. However, in terms of resilience, the traditional configuration remains superior in three out of the four examined scenarios.
This paper reviews and compares state-of-the-art cobalt-based catalysts and catalytic systems used to produce green and sustainable fuels using FTS. Being focused on comparing the effect of the catalyst formulation and synthesis method, the reactor type and operating parameters, as well as the quality of the obtained fuels, the aim is to identify the research gaps between these relevant research areas concerning production of green and sustainable fuels.
Mithilfe einer templatgestützten Synthese wurden poröse Kohlenstoffgerüste unter Verwendung von Silicagel als Templat hergestellt. Die chemische Gasphaseninfiltration (CVI) wurde hierbei als Synthese verwendet. Unter Variation verschiedener Reaktionsparameter zur Optimierung der Kohlenstoffabscheidung wurde dieser Prozess mathematisch modelliert and simuliert. Dabei konnten die experimentellen Ergebnisse gut mit den Modellen nachgebildet werden. Die zusätzliche Beschreibung der laminaren Strömung verbessert die Übereinstimmung deutlich.
Traditionally, sensitivity analysis has been utilized to determine the importance of input variables to a deep neural network (DNN). However, the quantification of sensitivity for each neuron in a network presents a significant challenge. In this article, a selective method for calculating neuron sensitivity in layers of neurons concerning network output is proposed. This approach incorporates scaling factors that facilitate the evaluation and comparison of neuron importance. Additionally, a hierarchical multi-scale optimization framework is proposed, where layers with high-importance neurons are selectively optimized. Unlike the traditional backpropagation method that optimizes the whole network at once, this alternative approach focuses on optimizing the more important layers. This paper provides fundamental theoretical analysis and motivating case study results for the proposed neural network treatment. The framework is shown to be effective in network optimization when applied to simulated and UCI Machine Learning Repository datasets. This alternative training generates local minima close to or even better than those obtained with the backpropagation method, utilizing the same starting points for comparative purposes within a multi-start optimization procedure. Moreover, the proposed approach is observed to be more efficient for large-scale DNNs. These results validate the proposed algorithmic framework as a rigorous and robust new optimization methodology for training (fitting) neural networks to input/output data series of any given system.
The Distributed Energy Systems (DES) or microgrid arose from the need to reduce greenhouse gases (GHG) emitted into the atmosphere by burning fossil fuels to generate energy. Reduction of energy losses, reconfiguration of the protection system and reduction of costs, and optimizing the configuration of these systems is recommended. Despite new research in literature, there is still a lack of optimization models that address the Brazilian reality. Therefore, the objective of this work is to introduce a decision-making framework for the design and operation of residential DES that takes into account the particularities of Brazil, based on mixed-integer programming models. The applicability of the framework is tested on a case study of a residential DES of 5 houses, located in Salvador, and used to compare scenarios pre- and post-COVID-19. The results show significant reduction in total annual cost and GHG emissions versus the base case without DES. This indicates that, although the country has a mostly “clean” energy matrix due to the use of hydroelectric plants, DES can enable improvement in residential electricity generation.