Chalmers Research
Not a member yet
88095 research outputs found
Sort by
Optimization of low-cost integration of wind and solar power in multi-node electricity systems: Mathematical modelling and dual solution approaches
Get PDFThe global production of electricity contributes significantly to the release of CO2 emissions. Therefore, a transformation of the electricity system is of vital importance in order to restrict global warming. This thesis concerns modelling and methodology of electricity systems which contain a large share of variable renewable electricity generation (i.e. wind and solar power).The two models developed in this thesis concern optimization of long-term investments in the electricity system. They aim at minimizing investment and production costs under electricity production constraints, using different spatial resolutions and technical detail, while meeting the electricity demand. These models are very large in nature due to the 1) high temporal resolution needed to capture the wind and solar variations while maintaining chronology in time, and 2) need to cover a large geographical scope in order to represent strategies to manage these variations (e.g.\ electricity trade). Thus, different decomposition methods are applied to reduce computation times. We develop three different decomposition methods: Lagrangian relaxation combined with variable splitting solved using either i) a subgradient algorithm or ii) an ADMM algorithm, and iii) a heuristic decomposition using a consensus algorithm. In all three cases, the decomposition is done with respect to the temporal resolution by dividing the year into 2-week periods. The decomposition methods are tested and evaluated for cases involving regions with different energy mixes and conditions for wind and solar power. Numerical results show faster computation times compared to the non-decomposed models and capacity investment options similar to the optimal solutions given by the latter models. However, the reduction in computation time may not be sufficient to motivate the increase in complexity and uncertainty of the decomposed models
Minimax-Bayes Reinforcement Learning
No full textWhile the Bayesian decision-theoretic framework offers an elegant solution to the problem of decision making under uncertainty, one question is how to appropriately select the prior distribution. One idea is to employ a worst-case prior. However, this is not as easy to specify in sequential decision making as in simple statistical estimation problems. This paper studies (sometimes approximate) minimax-Bayes solutions for various reinforcement learning problems to gain insights into the properties of the corresponding priors and policies. We find that while the worst-case prior depends on the setting, the corresponding minimax policies are more robust than those that assume a standard (i.e. uniform) prior
Synthesis, structure diversity, and antimicrobial studies of Ag(i) complexes with quinoline-type ligands
Get PDFCompounds [Ag(5NO2Qu)2]BF4 (1) and [Ag(Qu3CN)(H2O)]BF4 (2) were prepared and studied from a structural perspective and screened for antimicrobial activity. The Ag(i) in the monomeric complex 1 is coordinated to two 5-nitroquinoline (5NO2Qu) ligands via the N-atoms of the quinoline rings with equidistant Ag-N bonds (2.146(2) \uc5) and a N-Ag-N# bond angle of 171.42(8)\ub0. The 2D coordination polymer 2 contains tetracoordinated Ag(i) with two N-atoms (N1 and N2#1) from two quinoline-3-carbonitrile (Qu3CN) ligands and two O-atoms (O1 and O1#1) from two water molecules. The Qu3CN ligand acts as a connector between the Ag(i) sites along the b-direction via two short Ag1-N1 (2.185(4) \uc5) and Ag1-N2#1 (2.204(4) \uc5) bonds. In addition, the Ag(i) is coordinated with two symmetry related water molecules which are also acting as connectors between the Ag(i) sites along the a-direction via two longer Ag1-O1 (2.470(4) \uc5) and Ag1-O1#2 (2.546(4) \uc5) bonds. Hirshfeld surface analysis confirmed the significance of the polar F⋯H contacts in the molecular packing of 1 (25.9%) and 2 (39.9%). In addition, the crystal packing of 1 showed a significant amount of polar O⋯H (23.5%) contacts. Also, both complexes displayed π-π stacking interactions. The Ag(i) complexes and the free ligand were assessed for their antimicrobial activities. It was found that 1 (MIC = 7.8 μg mL−1) and 2 (MIC = 31.25 μg mL−1) have higher antifungal potency against C. albicans than their free ligands (MIC = 125 μg mL−1). Interestingly, 1 has better antifungal activity than the standard nystatin (15.6 μg mL−1). Also, both Ag(i) complexes and the free ligands as well have better activity against P. mirabilis than the common antibiotic amoxicillin
The EBLM project - IX. Five fully convective M-dwarfs, precisely measured with CHEOPS and TESS light curves
Get PDFEclipsing binaries are important benchmark objects to test and calibrate stellar structure and evolution models. This is especially true for binaries with a fully convective M-dwarf component for which direct measurements of these stars\u27 masses and radii are difficult using other techniques. Within the potential of M-dwarfs to be exoplanet host stars, the accuracy of theoretical predictions of their radius and effective temperature as a function of their mass is an active topic of discussion. Not only the parameters of transiting exoplanets but also the success of future atmospheric characterization relies on accurate theoretical predictions. We present the analysis of five eclipsing binaries with low-mass stellar companions out of a subsample of 23, for which we obtained ultra-high-precision light curves using the CHEOPS satellite. The observation of their primary and secondary eclipses are combined with spectroscopic measurements to precisely model the primary parameters and derive the M-dwarfs mass, radius, surface gravity, and effective temperature estimates using the PYCHEOPS data analysis software. Combining these results to the same set of parameters derived from TESS light curves, we find very good agreement (better than 1 percent for radius and better than 0.2 percent for surface gravity). We also analyse the importance of precise orbits from radial velocity measurements and find them to be crucial to derive M-dwarf radii in a regime below 5 percent accuracy. These results add five valuable data points to the mass-radius diagram of fully convective M-dwarfs
RAINBOW: Multi-Dimensional Hardware-Software Co-Design for DL Accelerator On-Chip Memory
No full textDeep Learning (DL) is developing at an extremely fast pace. The increased number of applications, optimizations and hardware devices available, results in a multi-dimensional design space where the best performance is achieved with a detailed analysis of the hardware-software co-design process. Furthermore, the high demands for memory and the off-chip latency cost result in the on-chip memory becoming critical for achieving high performance and efficiency. In this work, we propose RAINBOW, a tool to assist in the hardware-software co-design for DL accelerators\u27 on-chip memory. The purpose is to help the design and/or deployment of a DL model to a dedicated accelerator. RAINBOW generates different analyses results and feeds them to the optimizers. The result is a heterogeneous execution plan combining different approaches and techniques depending on the dynamic requirements and constraints. In our analysis, we concluded that given the opportunity, RAINBOW\u27S heterogeneous plans are able to reduce the DRAM accesses to approximately half when compared to homogeneous plans
Platinum-Based Nanocatalysts for Proton Exchange Membrane Fuel Cells
Get PDFFuel cells have potential to become an integral technology in a future sustainable energy system. For transport applications, the proton exchange membrane fuel cell (PEMFC) is the most promising option, exhibiting light weight and high energy density. However, large-scale commercialization is impeded by expensive catalyst materials and slow oxygen reduction reaction (ORR) kinetics on the cathode side. Several alternatives to the conventional platinum PEMFC catalyst have been proposed and studied during the last decades, one being platinum-rare earth (Pt-RE) metal alloys. With enhanced ORR activities and maintained stability, these materials are highly interesting for deployment in PEMFCs, and could potentially reduce both catalyst material use and overall fuel cell cost. In practical fuel cells, catalysts are required in nanoparticulate form, to facilitate sufficient performance while keeping material utilization high. Unfortunately, scalability remains as a main obstacle for Pt-RE nanoparticle synthesis, as fabrication of these materials has proven challenging, motivated by the high oxygen affinity of the rare-earth metals.This thesis investigates the use of sputtering onto liquid (SoL) substrates as a potential synthesis method for Pt-RE nanocatalysts. The influence of sputtering parameters, including substrate type and temperature, as well as gas environment, on the size and morphology of platinum-based nanocatalysts are studied. Transmission electron microscopy of platinum sputtered in four different liquids indicates that the size of the nanoparticles is only weakly dependent on temperature. Furthermore, catalyst layers fabricated from the SoL-synthesized nanocatalysts are evaluated in a half cell setup. The electrochemical results shows that high performing catalyst layer fabrication from SoL-synthesized nanoparticles is viable, which opens for further development of the technique
Accelerating CNN inference on long vector architectures via co-design
No full textCPU-based inference can be deployed as an alternative to off-chip accelerators. In this context, emerging vector architectures are a promising option, owing to their high efficiency. Yet the large design space of convolutional algorithms and hardware implementations makes the selection of design options challenging. In this paper, we present our ongoing research into co-designing future vector architectures for CPU-based Convolutional Neural Networks (CNN) inference focusing on the im2col+GEMM and Winograd kernels. Using the Gem5 simulator we explore the impact of several hardware microarchitectural features including (i) vector lanes, (ii) vector lengths, (iii) cache sizes, and (iv) options for integrating the vector unit into the CPU pipeline. In the context of im2col+GEMM, we study the impact of several BLIS-like algorithmic optimizations such as (1) utilization of vector registers, (2) loop unrolling, (3) loop reorder, (4) manual vectorization, (5) prefetching, and (6) packing of matrices, on the RISC-V Vector Extension and ARM-SVE ISAs. We use the YOLOv3 and VGG16 network models for our evaluation. Our co-design study shows that BLIS-like optimizations are not beneficial to all types of vector microarchitectures. We additionally demonstrate that longer vector lengths (of at least 8192 bits) and larger caches (of 256MB) can boost performance by 5
7, with our optimized CNN kernels, compared to a vector length of 512-bit and 1MB of L2 cache. In the context of Winograd, we present our novel approach of inter-tile parallelization across the input/output channels by using 8
78 tiles per channel to vectorize the algorithm on vector length agnostic (VLA) architectures. Our method exploits longer vector lengths and offers high memory reuse, resulting in performance improvement of up to 2.4
7 for non-strided convolutional layers with 3
73 kernel size, compared to our optimized im2col+GEMM approach on the Fujitsu A64FX processor. Our co-design study furthermore reveals that Winograd requires smaller cache sizes (up to 64MB) compared to im2col+GEMM
On the stabilization of emulsions by cellulose nanocrystals and nanofibrils: Interfacial behavior and synergism
Get PDFCompetitive adsorption of cellulose nanofibers at oil-water interface in Pickering emulsion is reported here. Dodecane-in-water emulsions stabilized by either of two types of nanocelluloses, cellulose nanocrystals (CNC) or cellulose nanofibrils (CNF), as well as by their binary mixtures with increasing fractions of CNC, were prepared using particle concentrations of 0.1–0.5 wt% and studied. Despite differences in shape and morphology, both forms of nanofibers produced stable emulsion droplets even at low particle concentrations (0.1 wt%), with CNC producing smaller droplets and emulsions with higher stability. When mixed, an increased fraction of CNC in the mixture reduced the average droplet size, which however applied only for higher contents of oil (30 and 50 wt%) and higher total contents of cellulose particles used under emulsification. The CNC particles controlled the size of emulsion droplets, while the role of CNF contributed to the further surface coverage. When the fraction of CNF in the mixture increased, the capability of CNC particles to readily adsorb at the oil-water interface was reduced by the CNF nanofibrils present in aqueous phase. The stability of emulsions with respect to changes in droplet size and creaming index was influenced more by oil content and total particle concentration than by the fraction of CNC present in the mixture
Filamentous fungi display different behavior during spruce bark degradation.
No full textMicrobial degradation of trees is hindered by their outermost tissue, the bark, which is also a material produced in vast amounts annually through debarking in the pulp and paper industry. While the bark is composed of the typical lignocellulose components cellulose, hemicellulose, and lignin, it also contains a large fraction of small molecules referred to as extractive compounds. The extractives are attributed to the strong anti-microbial properties of the bark, but despite their presence, filamentous fungi can frequently be found growing on the outside of fallen trees. How fungi deal with the presence of extractive compounds and polysaccharides while growing on bark is however not known today, and this precludes development of biological valorization methods. \ua0Here, we have followed fungi growing on spruce bark over six months, including white-rot (Dichomitus squalens), brown-rot (Postia placenta), as well as three Ascomycetes (Trichoderma reesei, Penicillium crustosum, Trichoderma sp. B1). The changes in the material were analyzed continuously using a combination of mass-loss determination, GC-MS, HPAEC-PAD, to monitor overall changes as well as detailed changes in extractives, carbohydrates, and lignin. These data enabled a comparison of the growth and substrate metabolism of five different fungi growing on the bark. The fungi exhibited clearly different approaches to the extractive compounds – from simply tolerating them, to detoxification and/or fully metabolizing them. Also carbohydrate analyses revealed significant differences among the fungi, with the brown-rot fungus P.\ua0 placenta displaying the typical hemicellulose first – cellulose second type of degradation pattern. From compositional analyses, D. squalens had the highest mass-loss (30%) and was quickest to reach stationary phase (12 weeks), and was able to significantly modify extractives and polysaccharides, in particular monosaccharides derived from pectin and xylan. Therefore, additional proteomic analyses was performed on D. squalens grown on bark, acetone-extracted bark (i.e. extractive-less), and galactomannan. The results revealed little difference in the proteome composition between acetone-extracted bark and bark, however, were identified. In particular, carbohydrate-active enzymes (CAZymes) related to pectin and xylan degradation were upregulated in the bark samples. \ua0Our results suggest that D. squalens, P. placenta, T. reesei, P. crustosum, Trichoderma sp. B1 have different substrate preferences, in particular, extractives are either degraded or tolerated. Significant changes could also be found in carbohydrate composition revealing pectin degradation. This work forms a basis for an understanding fungal degradation of bark
How generative adversarial networks promote the development of intelligent transportation systems: A survey
No full textIn current years, the improvement of deep learning has brought about tremendous changes: As a type of unsupervised deep learning algorithm, generative adversarial networks (GANs) have been widely employed in various fields including transportation. This paper reviews the development of GANs and their applications in the transportation domain. Specifically, many adopted GAN variants for autonomous driving are classified and demonstrated according to data generation, video trajectory prediction, and security of detection. To introduce GANs to traffic research, this review summarizes the related techniques for spatio-temporal, sparse data completion, and time-series data evaluation. GAN-based traffic anomaly inspections such as infrastructure detection and status monitoring are also assessed. Moreover, to promote further development of GANs in intelligent transportation systems (ITSs), challenges and noteworthy research directions on this topic are provided. In general, this survey summarizes 130 GAN-related references and provides comprehensive knowledge for scholars who desire to adopt GANs in their scientific works, especially transportation-related tasks