Chalmers Research
Not a member yet
88095 research outputs found
Sort by
Deep Learning-Based Connector Detection for Robotized Assembly of Automotive Wire Harnesses
No full textThe shift towards electrification and autonomous driving in the automotive industry results in more and more automotive wire harnesses being installed in modern automobiles, which stresses the great significance of guaranteeing the quality of automotive wire harness assembly. The mating of connectors is essential in the final assembly of automotive wire harnesses due to the importance of connectors on wire harness connection and signal transmission. However, the current manual operation of mating connectors leads to severe problems regarding assembly quality and ergonomics, where the robotized assembly has been considered, and different vision-based solutions have been proposed to facilitate a better perception of the robot control system on connectors. Nonetheless, there has been a lack of deep learning-based solutions for detecting automotive wire harness connectors in previous literature. This paper presents a deep learning-based connector detection for robotized automotive wire harness assembly. A dataset of twenty automotive wire harness connectors was created to train and evaluate a two-stage and a one-stage object detection model, respectively. The experiment results indicate the effectiveness of deep learning-based connector detection for automotive wire harness assembly but are limited by the design of the exteriors of connectors
Different electrodeposition techniques of manganese and nickel oxide on nickel foam and their effect on improved supercapacitor behaviour: a comparative study
No full textManganese and Nickel oxides were electrodeposited onto Nickel foam by potentiodynamic (10, 25 and 50 cycles), potentiostatic and galvanostatic modes and the effects of different electrodeposition techniques on the elemental compositions and their supercapactive behaviour were studied to optimise the most appropriate electrodeposition technique for supercapacitor application. The structural properties, morphology and elemental analysis were studied by X-ray Diffraction (XRD), Scanning Electron Microscopy (SEM) accompanied by Energy-Dispersive X-Ray Analysis (EDX). The electrodes’ functional groups were analysed via Fourier Transform-Infrared Spectroscopy (FT-IR). Their electrochemical supercapactive performance were assessed by calculating the areal capacitance from cyclic voltammograms (CV), from Galvanostatic charge-discharge Curves (GCD), and their behaviour was accessed by Electrochemical impedance spectroscopy (EIS) analyses in 0.1\ua0M KOH. The electrochemical results specified among the different electrode MN10, MN25, MN50 (potentiodynamic electrodeposition), MNCA (electrodeposition via chronoamperometry) and MNCP (electrodeposition via chronopotentiometry); MN25 delivered the highest areal capacitance areal capacitance 256.08\ua0F cm−2, with energy density 12.81 Wh cm−2 and power density 150.71\ua0W cm−2 with the capacitance retention percentage of 80.5% at 5 Acm−2 after 5000 cycles
LXL: LiDAR Excluded Lean 3D Object Detection with 4D Imaging Radar and Camera Fusion
No full textAs an emerging technology and a relatively affordable device, the 4D imaging radar has already been confirmed effective in performing 3D object detection in autonomous driving. Nevertheless, the sparsity and noisiness of 4D radar point clouds hinder further performance improvement, and in-depth studies about its fusion with other modalities are lacking. On the other hand, as a new image view transformation strategy, “sampling” has been applied in a few image-based detectors and shown to outperform the widely applied “depth-based splatting” proposed in Lift-Splat-Shoot (LSS), even without image depth prediction. However, the potential of “sampling” is not fully unleashed. This paper investigates the “sampling” view transformation strategy on the camera and 4D imaging radar fusion-based 3D object detection. LiDAR Excluded Lean (LXL) model, predicted image depth distribution maps and radar 3D occupancy grids are generated from image perspective view (PV) features and radar bird\u27s eye view (BEV) features, respectively. They are sent to the core of LXL, called “radar occupancy-assisted depth-based sampling”, to aid image view transformation. We demonstrated that more accurate view transformation can be performed by introducing image depths and radar information to enhance the “sampling” strategy. Experiments on VoD and TJ4DRadSet datasets show that the proposed method outperforms the state-of-the-art 3D object detection methods by a significant margin without bells and whistles. Ablation studies demonstrate that our method performs the best among different enhancement settings
A new open-source finite element lumbar spine model, its tuning and validation, and development of a tissue-based injury risk function for compression fractures
No full textLumbar spine fractures have been identified as a problem in motor vehicle crashes, and it is expected that this problem might increase with the introduction of reclined postures in autonomous vehicles. Human body models provide a means to address this issue and develop countermeasures. In this study a new open-source finite element lumbar spine model and an associated tissue-based injury risk function were developed and validated. The injury risk function was based on trabecular bone compressive strain in the superior-inferior direction. The kinematic and kinetic validation showed that the model compared reasonably to experimental data, with axial compression and flexion predictions being closest to experimental results. The new risk function was found to have a good quality index. Even though the model evaluations indicated that the fracture risk was somewhat overpredicted, it was judged that the current model, together with the associated injury risk function, can be used to estimate the risk for compressive fractures in the lumbar spine, with the knowledge that these estimates are most likely somewhat conservative
Combined model for 15N, 13C, and spin-state chemistry in molecular clouds
Get PDFWe present a new gas-grain chemical model for the combined isotopic fractionation of carbon and nitrogen in molecular clouds. To this end, we have developed gas-phase and grain-surface chemical networks where the isotope chemistry of carbon and nitrogen is coupled with a time-dependent description of spin-state chemistry, which is important for nitrogen chemistry at low temperatures. We updated the rate coefficients of some isotopic exchange reactions considered previously in the literature, and here we present a set of new exchange reactions involving molecules substituted in 13C and 15N simultaneously. We applied the model to a series of zero-dimensional simulations representing a set of physical conditions across a prototypical prestellar core, exploring the deviations of the isotopic abundance ratios in the various molecules from the elemental isotopic ratios as a function of physical conditions and time. We find that the 12C/13C ratio can deviate from the elemental ratio to a high factor depending on the molecule, and that there are highly time-dependent variations in the ratios. The HCN/H13CN ratio, for example, can obtain values of less than ten depending on the simulation time. The 14N/15N ratios tend to remain close to the assumed elemental ratio within approximately 10%, with no clearly discernible trends for the various species as a function of the physical conditions. Abundance ratios between 13C-containing molecules and 13C+15N-containing molecules however show somewhat increased levels of fractionation as a result of the newly included exchange reactions, though they still remain within a few tens of percent of the elemental 14N/15N ratio. Our results imply the existence of gradients in isotopic abundance ratios across prestellar cores, suggesting that detailed simulations are required to interpret observations of isotopically substituted molecules correctly, especially given that the various isotopic forms of a given molecule do not necessarily trace the same gas layers
Design and first application of a novel laboratory reactor for alkali studies in chemical looping applications
Get PDFAlkali compounds are readily released during biomass conversion and their complex interactions with reactor walls and sampling equipment makes detailed investigations challenging. This study evaluates a novel laboratory-scale fluidized bed reactor for chemical looping combustion (CLC) studies. The reactor design is based on detailed consideration of the behavior of alkali-containing molecules and aerosol particles and is guided by computational fluid dynamic simulations. The design allows for interactions between gaseous alkali and a fluidized bed, while minimizing alkali interactions with walls before and after the fluidized bed. The function of the laboratory reactor is demonstrated in experiments using online gas and alkali analysis. Alkali is continuously fed to the reactor as KOH or KCl aerosol with and without a fluidized bed of the oxygen carrier CaMn0.775Ti0.125Mg0.1O3-δ present in inert, reducing and oxidizing conditions at temperatures up to 900 \ub0C. Alkali uptake by the OC is characterized in all conditions, and observed to sensitively depend on gas composition, reactor temperature and type of alkali compound. The experimental setup is concluded to have a significantly improved functionality compared to a previously used reactor, which opens up for detailed studies of interactions between alkali compounds and oxygen carriers used in CLC
Digital transformation of the wood construction supply chain through building information modelling: Current state of practice
No full textPurpose: This study aims to present a state-of-the-art review of BIM in the Swedish construction practice with a focus on wood construction. It focuses on examining the extent, maturity, and actual practices of BIM in the Swedish wood construction industry, by analysing practitioners’ perspectives on the current state of BIM and its perceived benefits.Design/methodology/approach: A qualitative approach was selected, given the study\u27s exploratory character. Initially, an extensive review was undertaken to examine the current state of Building Information Modelling utilization and its associated advantages within the construction industry. Subsequently, empirical data was acquired through semi-structured interviews featuring open-ended questions, aimed at comprehensively assessing the prevailing extent of BIM integration within the Swedish wood construction sector.Findings: The research concluded that the wood construction industry in Sweden is shifting towards BIM on different levels, where in some cases, the level of implementation is still modest. It should be emphasised that the wood construction industry in Sweden is not realising the full potential of BIM. The industry is still using a combination of BIM and traditional methods, thus limiting the benefits that full BIM implementation could offer the industry.Originality: This study provided empirical evidence on the current perceptions and state of practice of the Swedish wood construction industry regarding BIM maturity
Impact of pole-to-pole DC voltage on energy requirement of FB YY-MMC
No full textThe aim of this paper is to investigate the impact of pole-to-pole DC voltage on the energy requirements of a modular multilevel converter with full-bridge submodules (FB-MMC). With focus on energy-storage equipped FB-MMCs, the relation between the pole-to-pole voltage and the peak-to-peak energy variations in the MMC\u27s arms is derived. Unlike MMC with half-bridge (HB) submodules (SM), in a FB-MMC the pole-to-pole DC voltage can be selected independently of the AC voltage; this allows for reduced energy requirements as well as lower number of FB SMs and capacitors for the same converter ratings. Furthermore, the impact of the converter\u27s circulating current on the MMC\u27s energy requirements is investigated. A guideline for the converter design is provided. Theoretical findings are validated via time-domain simulations
Developing Data Models for Smart Environmental Performance Management in Production
No full textFor manufacturing companies to prosper in the long term, they must demonstrate contribution to sustainable development by implementing greener practices using approaches such eco-efficiency and circular economy; i.e., creating social and economic value while minimising the environmental impact of production through efficient, closed-loop circulation of resources. In addition, industrial digitalization presents new opportunities to unlock new ways to measure complex systems’ performance and systematically improve towards circular economy and sustainability. This paper presents the results of a feasibility study aiming to develop a practical toolkit to implement environmental sustainability concepts at factory level. To achieve the project objective, we focused on data handling practices for environmental performance management, including process mapping, data inventory, data quality assessment, and gap analysis to identify existing strengths and define areas of improvement to boost the environmental performance of production systems
Linking ice and gas in the Coronet cluster in Corona Australis
Get PDFContext. During the journey from the cloud to the disc, the chemical composition of the protostellar envelope material can be either preserved or processed to varying degrees depending on the surrounding physical environment. Aims. This works aims to constrain the interplay of solid (ice) and gaseous methanol (CH3OH) in the outer regions of protostellar envelopes located in the Coronet cluster in Corona Australis (CrA), and assess the importance of irradiation by the Herbig Ae/Be star R CrA. CH3OH is a prime test case as it predominantly forms as a consequence of the solid-gas interplay (hydrogenation of condensed CO molecules onto the grain surfaces) and it plays an important role in future complex molecular processing. Methods. We present 1.3 mm Submillimeter Array (SMA) and Atacama Pathfinder Experiment (APEX) observations towards the envelopes of four low-mass protostars in the Coronet cluster. Eighteen molecular transitions of seven species were identified. We calculated CH3OH gas-to-ice ratios in this strongly irradiated cluster and compared them with ratios determined towards protostars located in less irradiated regions such as Serpens SVS 4 in Serpens Main and the Barnard 35A cloud in the λ Orionis region. Results. The CH3OH gas-to-ice ratios in the Coronet cluster vary by one order of magnitude (from 1.2
7 10-4 to 3.1
7 10-3) which is similar to less irradiated regions as found in previous studies. We find that the CH3OH gas-to-ice ratios estimated in these three regions are remarkably similar despite the different UV radiation field intensities and formation histories. Conclusions. This result suggests that the overall CH3OH chemistry in the outer regions of low-mass envelopes is relatively independent of variations in the physical conditions and hence that it is set during the prestellar stage