FB Umweltplanung/-technik (UCB)
Filtern
Erscheinungsjahr
Dokumenttyp
Volltext vorhanden
- ja (44)
Gehört zur Bibliographie
- nein (44)
Schlagworte
- Elektrofahrzeug (5)
- Biodiversität (4)
- CO2-Bilanz (4)
- Elektromobilität (4)
- Fermentation (4)
- Maschinelles Lernen (4)
- Biomonitoring (3)
- Insekten (3)
- Klimaänderung (3)
- Rapid Prototyping <Fertigung> (3)
Institut
Numerous research methods have been developed to detect anomalies in the areas of security and risk analysis. In healthcare, there are numerous use cases where anomaly detection is relevant. For example, early detection of sepsis is one such use case. Early treatment of sepsis is cost effective and reduces the number of hospital days of patients in the ICU. There is no single procedure that is sufficient for sepsis diagnosis, and combinations of approaches are needed. Detecting anomalies in patient time series data could help speed the development of some decisions. However, our algorithm must be viewed as complementary to other approaches based on laboratory values and physician judgments. The focus of this work is to develop a hybrid method for detecting anomalies that occur, for example, in multidimensional medical signals, sensor signals, or other time series in business and nature. The novelty of our approach lies in the extension and combination of existing approaches: Statistics, Self Organizing Maps and Linear Discriminant Analysis in a unique and unprecedented way with the goal of identifying different types of anomalies in real-time measurement data and defining the point where the anomaly occurs. The proposed algorithm not only has the full potential to detect anomalies, but also to find real points where an anomaly starts.
Background: In recent years, the volume of medical knowledge and health data has increased rapidly. For example, the increased availability of electronic health records (EHRs) provides accurate, up-to-date, and complete information about patients at the point of care and enables medical staff to have quick access to patient records for more coordinated and efficient care. With this increase in knowledge, the complexity of accurate, evidence-based medicine tends to grow all the time. Health care workers must deal with an increasing amount of data and documentation. Meanwhile, relevant patient data are frequently overshadowed by a layer of less relevant data, causing medical staff to often miss important values or abnormal trends and their importance to the progression of the patient’s case.
Objective: The goal of this work is to analyze the current laboratory results for patients in the intensive care unit (ICU) and classify which of these lab values could be abnormal the next time the test is done. Detecting near-future abnormalities can be useful to support clinicians in their decision-making process in the ICU by drawing their attention to the important values and focus on future lab testing, saving them both time and money. Additionally, it will give doctors more time to spend with patients, rather than skimming through a long list of lab values.
Methods: We used Structured Query Language to extract 25 lab values for mechanically ventilated patients in the ICU from the MIMIC-III and eICU data sets. Additionally, we applied time-windowed sampling and holding, and a support vector machine to fill in the missing values in the sparse time series, as well as the Tukey range to detect and delete anomalies. Then, we used the data to train 4 deep learning models for time series classification, as well as a gradient boosting–based algorithm and compared their performance on both data sets.
Results: The models tested in this work (deep neural networks and gradient boosting), combined with the preprocessing pipeline, achieved an accuracy of at least 80% on the multilabel classification task. Moreover, the model based on the multiple convolutional neural network outperformed the other algorithms on both data sets, with the accuracy exceeding 89%.
Conclusions: In this work, we show that using machine learning and deep neural networks to predict near-future abnormalities in lab values can achieve satisfactory results. Our system was trained, validated, and tested on 2 well-known data sets to ensure that our system bridged the reality gap as much as possible. Finally, the model can be used in combination with our preprocessing pipeline on real-life EHRs to improve patients’ diagnosis and treatment.
The number of additive manufacturing methods and materials is growing rapidly, leaving gaps in the knowledge of specific material properties. A relatively recent addition is the metal-filled filament to be printed similarly to the fused filament fabrication (FFF) technology used for plastic materials, but with additional debinding and sintering steps. While tensile, bending, and shear properties of metals manufactured this way have been studied thoroughly, their fatigue properties remain unexplored. Thus, the paper aims to determine the tensile, fatigue, and impact strengths of Markforged 17-4 PH and BASF Ultrafuse 316L stainless steel to answer whether the metal FFF can be used for structural parts safely with the current state of technology. They are compared to two 316L variants manufactured via selective laser melting (SLM) and literature results. For extrusion-based additive manufacturing methods, a significant decrease in tensile and fatigue strength is observed compared to specimens manufactured via SLM. Defects created during the extrusion and by the pathing scheme, causing a rough surface and internal voids to act as local stress risers, handle the strength decrease. The findings cast doubt on whether the metal FFF technique can be safely used for structural components; therefore, further developments are needed to reduce internal material defects.
1. Recent reports on insect decline have highlighted the need for long-term data on insect communities towards identifying their trends and drivers.
2. With the launch of many new insect monitoring schemes to investigate insect communities over large spatial and temporal scales, Malaise traps have become one of the most important tools due to the broad spectrum of species collected and reduced capture bias through passive sampling of insects day and night. However, Malaise traps can vary in size, shape, and colour, and it is unknown how these differences affect biomass, species richness, and composition of trap catch, making it difficult to compare results between studies.
3. We compared five Malaise trap types (three variations of the Townes and two variations of the Bartak Malaise trap) to determine their effects on biomass and species richness as identified by metabarcoding.
4. Insect biomass varied by 20%–55%, not strictly following trap size but varying with trap type. Total species richness was 20%–38% higher in the three Townes trap models compared to the Bartak traps. Bartak traps captured lower richness of highly mobile taxa but increased richness of ground-dwelling taxa. The white roofed Townes trap captured a higher richness of pollinators.
5. We find that biomass, total richness, and taxa group specific richness are all sensitive to Malaise trap type. Trap type should be carefully considered and aligned to match monitoring and research questions. Additionally, our estimates of trap type effects can be used to adjust results to facilitate comparisons across studies.
While the contribution of renewable energy technologies to the energy system is increasing, so is its level of complexity. In addition to new types of consumer systems, the future system will be characterized by volatile generation plants that will require storage technologies. Furthermore, a solid interconnected system that enables the transit of electrical energy can reduce the need for generation and storage systems. Therefore, appropriate methods are needed to analyze energy production and consumption interactions within different system constellations. Energy system models can help to understand and build these future energy systems. However, although various energy models already exist, none of them can cover all issues related to integrating renewable energy systems. The existing research gap is also reflected in the fact that current models cannot model the entire energy system for very high shares of renewable energies with high temporal resolution (15 min or 1-h steps) and high spatial resolution. Additionally, the low availability of open-source energy models leads to a lack of transparency about exactly how they work. To close this gap, the sector-coupled energy model (UCB-SEnMod) was developed. Its unique features are the modular structure, high flexibility, and applicability, enabling it to model any system constellation and can be easily extended with new functions due to its software design. Due to the software architecture, it is possible to map individual buildings or companies and regions, or even countries. In addition, we plan to make the energy model UCB-SEnMod available as an open-source framework to enable users to understand the functionality and configuration options more easily. This paper presents the methodology of the UCB-SEnMod model. The main components of the model are described in detail, i.e., the energy generation systems, the consumption components in the electricity, heat, and transport sectors, and the possibilities of load balancing.
This paper presents a feasibility study for the production of recycled glycol modified polyethylene terephthalate (PETG) material for additive manufacturing. Past studies showed a variety of results for the recycling of 3D-printing material, therefore the precise effect on the material properties is not completely clear. For this work, PETG waste of the same grade was recycled once and further processed into 3D printing filament. The study compares three blend ratios between purchased plastic pellets and recycled pellets to determine the degradation effect of one recycling cycle and possible blend ratios to counter these effects. Furthermore, the results include a commercially available filament. The comparison uses the filament diameter, the dimensional accuracy of the printed test specimen and mechanical properties as quality criteria. The study shows that the recycled material has a minor decrease concerning the tensile strength and Young’s modulus.
Research in global change ecology relies heavily on global climatic grids derived from estimates of air temperature in open areas at around 2 m above the ground. These climatic grids do not reflect conditions below vegetation canopies and near the ground surface, where critical ecosystem functions occur and most terrestrial species reside. Here, we provide global maps of soil temperature and bioclimatic variables at a 1-km2 resolution for 0–5 and 5–15 cm soil depth. These maps were created by calculating the difference (i.e. offset) between in situ soil temperature measurements, based on time series from over 1200 1-km2 pixels (summarized from 8519 unique temperature sensors) across all the world's major terrestrial biomes, and coarse-grained air temperature estimates from ERA5-Land (an atmospheric reanalysis by the European Centre for Medium-Range Weather Forecasts). We show that mean annual soil temperature differs markedly from the corresponding gridded air temperature, by up to 10°C (mean = 3.0 ± 2.1°C), with substantial variation across biomes and seasons. Over the year, soils in cold and/or dry biomes are substantially warmer (+3.6 ± 2.3°C) than gridded air temperature, whereas soils in warm and humid environments are on average slightly cooler (−0.7 ± 2.3°C). The observed substantial and biome-specific offsets emphasize that the projected impacts of climate and climate change on near-surface biodiversity and ecosystem functioning are inaccurately assessed when air rather than soil temperature is used, especially in cold environments. The global soil-related bioclimatic variables provided here are an important step forward for any application in ecology and related disciplines. Nevertheless, we highlight the need to fill remaining geographic gaps by collecting more in situ measurements of microclimate conditions to further enhance the spatiotemporal resolution of global soil temperature products for ecological applications.
Carbon footprinting of universities worldwide: Part I — objective comparison by standardized metrics
(2021)
Background: Universities, as innovation drivers in science and technology worldwide, should be leading the Great Transformation towards a carbon–neutral society and many have indeed picked up the challenge. However, only a small number of universities worldwide are collecting and publishing their carbon footprints, and some of them have defined zero emission targets. Unfortunately, there is limited consistency between the reported carbon footprints (CFs) because of different analysis methods, different impact measures, and different target definitions by the respective universities.
Results: Comprehensive CF data of 20 universities from around the globe were collected and analysed. Essential factors contributing to the university CF were identified. For the first time, CF data from universities were not only compared. The CF data were also evaluated, partly corrected, and augmented by missing contributions, to improve the consistency and comparability. The CF performance of each university in the respective year is thus homogenized, and measured by means of two metrics: CO2e emissions per capita and per m2 of constructed area. Both metrics vary by one order of magnitude across the different universities in this study. However, we identified ten universities reaching a per capita carbon footprint of lower than or close to 1.0 Mt (metric tons) CO2e/person and year (normalized by the number of people associated with the university), independent from the university’s size. In addition to the aforementioned two metrics, we suggested a new metric expressing the economic efficiency in terms of the CF per $ expenditures and year. We next aggregated the results for all three impact measures, arriving at an overall carbon performance for the respective universities, which we found to be independent of geographical latitude. Instead the per capita measure correlates with the national per capita CFs, and it reaches on average 23% of the national impacts per capita. The three top performing universities are located in Switzerland, Chile, and Germany.
Conclusion: The usual reporting of CO2 emissions is categorized into Scopes 1–3 following the GHG Protocol Corporate Accounting Standard which makes comparison across universities challenging. In this study, we attempted to standardize the CF metrics, allowing us to objectively compare the CF at several universities. From this study, we observed that, almost 30 years after the Earth Summit in Rio de Janeiro (1992), the results are still limited. Only one zero emission university was identified, and hence, the transformation should speed up globally.
Laboratory protocols using magnetic beads have gained importance in the purification of mRNA for vaccines. Here, the produced mRNA hybridizes specifically to oligo(dT)-functionalized magnetic beads after cell lysis. The mRNA-loaded magnetic beads can be selectively separated using a magnet. Subsequently, impurities are removed by washing steps and the mRNA is eluted. Magnetic separation is utilized in each step, using different buffers such as the lysis/binding buffer. To reduce the time required for purification of larger amounts of mRNA vaccine for clinical trials, high-gradient magnetic separation (HGMS) is suitable. Thereby, magnetic beads are selectively retained in a flow-through separation chamber. To meet the requirements of biopharmaceutical production, a disposable HGMS separation chamber with a certified material (United States Pharmacopeia Class VI) was developed which can be manufactured using 3D printing. Due to the special design, the filter matrix itself is not in contact with the product. The separation chamber was tested with suspensions of oligo(dT)-functionalized Dynabeads MyOne loaded with synthetic mRNA. At a concentration of cB = 1.6–2.1 g·L–1 in lysis/binding buffer, these 1 μm magnetic particles are retained to more than 99.39% at volumetric flows of up to 150 mL·min–1 with the developed SU-HGMS separation chamber. When using the separation chamber with volumetric flow rates below 50 mL·min–1, the retained particle mass is even more than 99.99%.
Purification of mRNA with oligo(dT)-functionalized magnetic particles involves a series of magnetic separations for buffer exchange and washing. Magnetic particles interact and agglomerate with each other when a magnetic field is applied, which can result in a decreased total surface area and thus a decreased yield of mRNA. In addition, agglomeration may also be caused by mRNA loading on the magnetic particles. Therefore, it is of interest how the individual steps of magnetic separation and subsequent redispersion in the buffers used affect the particle size distribution. The lysis/binding buffer is the most important buffer for the separation of mRNA from the multicomponent suspension of cell lysate. Therefore, monodisperse magnetic particles loaded with mRNA were dispersed in the lysis/binding buffer and in the reference system deionized water, and the particle size distributions were measured. A concentration-dependent agglomeration tendency was observed in deionized water. In contrast, no significant agglomeration was detected in the lysis/binding buffer. With regard to magnetic particle recycling, the influence of different storage and drying processes on particle size distribution was investigated. Agglomeration occurred in all process alternatives. For de-agglomeration, ultrasonic treatment was examined. It represents a suitable method for reproducible restoration of the original particle size distribution.