arXiv subjects

V. Breschi

Publications and source records attributed to V. Breschi.

5 recordsLinked to original sources

arXiv2026

The impact of sensor placement on graph-neural-network-based leakage detection

Sensor placement for leakage detection in water distribution networks is an important and practical challenge for water utilities. Recent work has shown that graph neural networks can estimate and predict pressures and detect leaks, but their performance strongly depends on the available sensor measurements and configurations. In this paper, we investigate how sensor placement influences the performance of GNN-based leakage detection. We propose a novel PageRank-Centrality-based sensor placement method and demonstrate that it substantially impacts reconstruction, prediction, and leakage detection on the EPANET Net1.

cs.LG↗

arXiv2025

Scalable Sensor Placement for Cyclic Networks with Observability Guarantees: Application to Water Distribution Networks

Optimal sensor placement is essential for state estimation and effective network monitoring. As known in the literature, this problem becomes particularly challenging in large-scale undirected or bidirected cyclic networks with parametric uncertainties, such as water distribution networks (WDNs), where pipe resistance and demand patterns are often unknown. Motivated by the challenges of cycles, parametric uncertainties, and scalability, this paper proposes a sensor placement algorithm that guarantees structural observability for cyclic and acyclic networks with parametric uncertainties. By leveraging a graph-based strategy, the proposed method efficiently addresses the computational complexities of large-scale networks. To demonstrate the algorithm's effectiveness, we apply it to several EPANET benchmark WDNs. Most notably, the developed algorithm solves the sensor placement problem with guaranteed structured observability for the L-town WDN with 1694 nodes and 124 cycles in under 0.1 seconds.

eess.SY↗

arXiv2024

Exploiting structural observability and graph colorability for optimal sensor placement in water distribution networks

Water distribution networks (WDNs) are critical systems for our society and detecting leakages is important for minimizing losses and water waste. This makes optimal sensor placement for leakage detection very relevant. Existing sensor placement methods rely on simulation-based scenarios, often lacking structure and generalizability, or depend on the knowledge of specific parameters of the WDN as well as on initial sensor data for linearization and demand estimation. Motivated by this, this paper investigates the observability of an entire WDN, based on structural observability theory. This allows us to establish the conditions for the observability of the WDN model, independently of parameter uncertainties. Additionally, a sensor placement algorithm is proposed that leverages such observability conditions and graph theory and accounts for the industrial and material costs. To demonstrate the effectiveness of our approach, we apply it to a hydraulic-transient WDN model.

eess.SY↗

arXiv2020

Social network analysis of electric vehicles adoption: a data-based approach

Mobility is undergoing dramatic transformations. Especially in the context of urban areas, several significant changes are underway, driven by both new mobility needs and environmental concerns. The most mature one, which still is struggling to affirm itself is the process of the adoption of Electric Vehicles (EVs), thus switching from fuel-based to battery-powered propulsion technologies. Many social and economic barriers have proved to play a crucial role in this process, ranging from level of education, environmental awareness, age and census. This work aims at contributing to the study of this adoption process through a data-based lens, using real mobility patterns to setup a social-network analysis to model the spread of consensus among neighbouring people that can enable the switch to EVs. In particular, we build the network topology using proximity measures that emerge from the analysis of real trips, and the initial disposition of the single agents towards the EV technology is inferred from their real mobility patterns. Based on this network, a cascade adoption model is simulated to investigate the dynamics of the adoption process, and an incentive scheme is designed to show how different policies can contribute to the opinion diffusion over time on the network.

cs.SI↗

arXiv2017

Fitting Jump Models

We describe a new framework for fitting jump models to a sequence of data. The key idea is to alternate between minimizing a loss function to fit multiple model parameters, and minimizing a discrete loss function to determine which set of model parameters is active at each data point. The framework is quite general and encompasses popular classes of models, such as hidden Markov models and piecewise affine models. The shape of the chosen loss functions to minimize determine the shape of the resulting jump model.

cs.LG↗