Felix Schmid

Accurate real-time flood forecasting under critical infrastructure failure remains a challenge in urban flood risk management. This study presents a real-time data-driven forecast system for spatiotemporal prediction of flood inundation resulting from stormwater pump station failures in a polder-type urban catchment in the Ruhr region, Germany. A Convolutional Neural Network (CNN) based on a feature-informed residual network is developed to surrogate a 1D/2D coupled hydrodynamic model, enabling rapid and dynamic prediction of maximum water depths across all significant pump failure scenarios for the next 24 hours. To accommodate long-duration events exceeding the prediction window and to avoid the error accumulation in autoregressive architectures, a computationally efficient dynamic windowing approach with a time step of 6 h is implemented, integrating accumulated rainfall and discharge as temporal memory. Sub-classified probabilistic flood maps (PFMs) are subsequently derived through a likelihood-weighted approach, combining scenario-specific predictions with pump failure probabilities estimated via a Bayesian network. Evaluated on 56 test events representing long-term measurements, synthetic design storms, and disaggregated climate scenarios, the model achieves mean RMSE values between 0.02 m and 0.04 m and Critical Success Index values consistently exceeding 0.95 across all failure scenarios. The proposed framework enables operational real-time assessment of spatially distributed flood hazard under pump failure conditions, supporting targeted emergency response and flood risk management decisions.

Urban mobility is increasingly becoming accepted as a basic human need, as socio-economic opportunities depend on the ability to reach places within an acceptable time. Conversely, the emergence of megalopoleis as dominant features of the global landscape has increased commuting effort to unprecedented levels, due to the ever expanding urban areas and the associated travel distances. This now poses a risk to the efficient accessibility of cities, but there is an assumption that the problem can be overcome by increasing the speed of transport systems. However, advocates of this approach overlook important utility trade-offs that arise from the conflict between greater vehicle speeds and the additional time required to access the services. In this paper, we investigate this approach and show that higher speeds in metro systems do not always result in faster travel in cities. We then propose a new approach to addressing the problem, which culminates in a solution that can overcome the current paradoxes and increase door-to-door speeds more effectively. The resulting operational concept optimizes speed and coverage in urban rail systems in megalopoleis, accommodating the longer trips within time budgets. We position this research as a starting point to a new perspective on developing complex urban systems in the future.

The railway control system is designed to act as a safety system that prevents collisions between trains and that maximises the utilisation of the rail network. Different signalling systems are widely used to achieve safety, but to achieve higher capacity, these existing systems may have to be updated or replaced in the future. Combining moving block and aspects of relative distance braking, a novel signalling system is proposed in this paper. The simulation shows that the novel concept has an excellent capacity outcome. The safety risk is also discussed to ensure its feasibility.

Models have the ability to represent a specific viewpoint of situations or circumstances in the real world. This paper facilitates the understanding of the relationship between railway safety and operational performance by providing an overview of performance models and presenting a model framework that incorporates both elements of safety and operational performance in the rail industry. This framework, if further developed into a working model with complex equations will facilitate decision making by governments and stakeholders in the daily operations and investments of railways.

Platform screen doors (PSD) are the sliding barrier doors installed at the edges of station platforms in many modern metro stations and occasionally on heavy rail systems. They serve many functions among which are the suicide prevention, optimisation of station energy consumption and safety, particularly by shielding passengers from gaining access to the rail tracks. PSD can prolong dwell time, increase the danger of mantrap and can extend emergency evacuation periods, among others. These pave way for the existence of PSD to have both positive and negative effects on the railway system. This research identified the benefits that could be derived from installing PSD at train stations and the drawbacks associated with its presence, with a view of evaluating the overall effects that the PSD has on the railway system. Upon identification of both factors, a causal loop diagram (CLD) has been developed to enable a pictorial demonstration of the various PSD effects, their systematic interrelations, the polarity of effects, and how they combine to affect the entire railway system. The conclusion reached in the present study was that PSD as a system can have significant impact on many aspects of the railway such as delays, capacity, safety, platform climate control, energy consumption, air quality and so on. But there are no studies that attempt to combine them all into an overall evaluation.

Driving style has emerged as an important determinant of fuel economy. There is now evidence that driving style can be influenced to improve fuel economy as well as other aspects such as safety. However, it is not clear which are the most appropriate and influential factors that affect an individual's, or a group's, driving style with respect to improving fuel economy. In this paper, such factors were identified from the literature and by way of driver training programmes for fuel economy. The factors were then categorised under driver factors, operating the vehicle, vehicle dynamics and driver awareness. The influences of the factors on fuel economy were prioritised using a multi-criteria analysis method called the analytical hierarchy process to determine their relative importance. It was found that driver awareness, measured in terms of culture change and better management, was considered the most influential category. The second most influential category of factors concerned operating the vehicle or vehicle control where acceleration and speed were dominant. The driver-related factors were considered to have the least influence on fuel economy. The results can be used to improve interventions such as driver training by emphasising the most influential driving factors.