Repository for Publications and Research Data
News from the ETH Library
Recently Added
Statistical integro-differential fracture model (Sid-FM) for homogenised scalar transport
Item type: Journal Article
Cao, Shangyi; Stalder, Daniel; Meyer, Daniel W.; et al. (2026)
This work presents an efficient statistical model to simulate expected scalar transport in fractured porous media below the representative elementary volume scale. We focus on embedded, highly conductive, isolated fractures. The statistical integro-differential fracture model (Sid-FM) solves for ensemble-averaged solutions directly, avoiding computationally expensive Monte Carlo simulation. The expected fluid exchange between isolated fractures and the porous matrix is modelled via a non-local kernel function, leading to a set of integro-differential equations. The model is validated against reference data from Monte Carlo simulations for statistically one-dimensional test cases and shows good agreement.
High-resolution modelling of organic aerosol over Europe: exploring spatial and temporal variability and drivers
Item type: Journal Article
Banos, Daniel Trejo; Upadhyay, Abhishek; Cheng, Yun; et al. (2026)
Organic aerosol (OA) is a major component of atmospheric particulate matter (PM), affecting both human health and climate. However, high-resolution estimates of OA exposure needed for exposure analysis remain scarce. Here, we integrate a chemical transport model (CAMx) with a random forest (RF) machine learning approach to bias-correct and downscale daily OA concentrations across Europe. CAMx OA simulations at ∼15 km resolution show moderate agreement with observations (r = 0.55). By combining these outputs with high-resolution land-use data and training the RF model on ∼48,000 daily OA measurements from 137 sites, prediction accuracy improved (r = 0.65), with ∼l5% reduction in root mean square error. The resulting maps provide European daily OA concentrations at ∼250 m resolution for alternate years from 2011 to 2019. The model captures key spatial features, including elevated OA in the Po Valley, Southeastern, and Central Europe, as well as intracity variations due to local hotspots. Seasonal analysis reveals higher concentrations in winter, while long-term trends indicate a general decline in OA levels. Exposure estimates show that half of the European population experiences OA levels above 3 µg/m3, and ∼50 million people are exposed to more than 5 µg/m3, which is the current guideline level recommended by the world health organization for total PM2.5. These high-resolution OA maps offer vital critical support for epidemiological research and air quality policy.
Deep Learning-Based Computer Vision for Automated Wood Grading: Enabling Efficient and Adaptive Process Optimization
Item type: Doctoral Thesis
Achatz, Julia Magdalena (2026)
Wood is a natural, renewable material with excellent aesthetic and mechanical properties. Its increased use in construction can also support climate-change mitigation by acting as a long-term carbon store. However, industrial processing is challenging due to wood’s inherent heterogeneity. Climate change intensifies this problem by shifting forest composition toward a higher proportion of hardwoods, which increase heterogeneity even more. To address this, effective wood sorting is key, as it manages heterogeneity, boosts processing efficiency, and maximizes material use. At the same time, current automated sorting solutions remain costly and often difficult to integrate into existing processing lines due to space constraints. Consequently, many small sawmills still sort wood manually or not at all, underscoring the need for more accessible approaches. This thesis develops automated wood sorting systems using deep learning–based computer vision, providing a lightweight, cost-efficient and flexible alternative to existing market solutions. It covers multiple stages of the wood value chain, focusing on three sorting tasks: roundwood, lamella, and rod sorting.
The first project starts at an early stage of the wood value chain, focusing on sorting roundwood by species and quality. By installing a camera system on an industrial sorting line, we showed that end-to-end Convolutional Neural Networks (CNNs) can effectively predict species and quality from visual data.
However, deep neural networks are considered black boxes and are difficult for humans to interpret, which limits their industrial acceptance. To address this issue, we employed Explainable Artificial Intelligence (XAI) methods to evaluate model robustness. Furthermore, we developed the Segmentation Decision Tree (SegDT) approach. A novel concept of combining a You Only Look Once (YOLO)v8x instance segmentation model to detect defects with a decision tree classifier to predict the quality. This hybrid-interpretable method matches the predictive accuracy of end-to-end CNNs for quality assessment, while providing defect-level insights.
Beyond species and surface quality, mechanical properties are crucial for construction. The second project moves further along the wood value chain, focusing on lamella. We captured lamella images with an industrial line-scan camera and labeled them with four-point bending test results. End-to-end models and an extended SegDT approach effectively predicted both quality and mechanical properties from these images.
So far, we focused on standardized industrial processes with fixed dimensions, established techniques, and widely used species. The final project goes beyond current capabilities by sorting split wooden rods for novel hardwood boards. These rods have irregular shapes and variable dimensions, making the task significantly more challenging. Models predicted species reliably, while estimating mechanical properties remained challenging. Although deep neural networks extracted valuable information from the images, much of the variation remained unexplained.
This thesis shows that deep learning–based computer vision offers significant potential for cost-efficient wood sorting. The approach is flexible and affordable, suitable even for small companies, and can be deployed on mobile devices, for example for in-forest sorting of logs. By lowering production costs and improving material yield, it can strengthen the role of wood in construction.
From Cognitive Overload to Optimization of electricity self-consumption: The Impact of a New Digital Tool
Item type: Working Paper
Filippini, Massimo; Srinivasan, Suchita; Sureka, Keshav (2026)
In this paper, we investigate how an automated home energy management system (AHEMS), operated through a new app can help consumers optimize their decisions. Optimizing self-consumption is a complex task for prosumers, i.e., households that produce, consume, and sell electricity to the grid, requiring cognitive effort. We study the impact of a novel app that assists prosumers in optimizing the self-consumption by automating the operation of electrical appliance and heating system. The main goals of this paper are twofold: first, to analyze the impact of the electricity price shock resulting from the war in Ukraine in 2022 on the adoption of this app, and secondly, to examine how the adoption of this AHEMS affected self-consumption levels. For this analysis, we employ a Difference-in-Differences model and an IV model using data from approximately 1,200 prosumers in Switzerland, observed in 2022 and 2023. The empirical results indicate that prosumers facing a significant increase in electricity prices increased their adoption of specialized subscriptions that facilitate automated management of electrical appliances by approximately 5%. Furthermore, prosumers with such subscriptions experience 9-11 percentage points higher self-consumption rates than non-subscribers. For a representative prosumer of our sample, this increase in self-consumption provides monetary savings that exceed the cost of the subscription.
Social insights into sustainable adoption of shared autonomous vehicles
Item type: Journal Article
Israel, Fabian; Heinen, Eva; Plaut, Pnina (2026)
The adoption of autonomous vehicles (AVs) is expected to reshape travel behaviors and spatial patterns, with significant societal implications. In this transition, individual and social values are critical for the sustainable uptake of shared autonomous vehicles (SAVs), particularly when trips are shared (i.e., pooled) among multiple users. However, strong dependence on and emotional attachment to private cars remain key barriers to adopting shared mobility systems. This study examines how generational values, social factors, and travel habits influence willingness to adopt SAVs for shared trips—an essential component of sustainable transport transitions aimed at inclusivity, reduced emissions, and efficient resource use. Adopting a qualitative exploratory approach, four focus groups conducted in the Tel Aviv metropolitan area reveal how life-stage experiences and social values shape openness to shared automated mobility. The findings highlight how social considerations, personal preferences, and habitual travel routines affect willingness to adopt shared SAV trips. Key insights include: (1) current travel experiences and emotional attachments shape openness to future SAV use; (2) gaps between public discourse and user perceptions underscore the need for clearer communication; (3) personalized ridesharing enhances comfort while raising equity concerns; and (4) generational differences suggest that SAV systems must adapt to diverse mobility needs across life stages. While based on a small, urban, and relatively homogeneous sample, the study offers first-order insights that advance understanding of social acceptance and inform a sustainable and inclusive SAV research and implementation agenda.