1,721,040 research outputs found
SmartH2O: an integrated platform coupling smart water meters with ICT and data intensive modeling to support residential water management
No full textThe design of user-customized water demand management strategies (WDMS) does require detailed information on the users’ water consumption habits and demand patterns. While low-resolution consumption data, as traditionally collected for billing purposes, do not allow for detailed investigations on consumers’ behavior, the recent advent of smart water sensors allows metering residential water consumption with sub-daily frequency. High-frequency water data are advancing our ability in characterizing and modeling consumers’ behavior, thus supporting the design of user-oriented residential water demand management strategies. Furthermore, they allow to close the information and communication loop between water utilities and their customers, as they provide instantaneous information on the state of the network and allow to continuously inform the users on their consumption and savings.
SmartH2O is a EU funded project which develops an ICT platform for raising users’ awareness about water consumption and pursuing water savings in the residential sector thanks to the integrated use of smart meters, advanced behavioral modeling, social computation, and personalized demand management strategies such as customized feedbacks and price schemes.
In this work, we present the main components of the SmartH2O integrated platform, with a particular focus on (i) high-resolution water consumption data collection through smart meters, (ii) users’ consumption modeling at the house level and end-use characterization through automatic disaggregation algorithms, and (iii) water consumption data visualization
Combining wavelet-enhanced feature selection and deep learning techniques for multi-step forecasting of urban water demand
Full text linkUrban water demand (UWD) forecasting is essential for water supply network optimization and management, both in business-as-usual scenarios, as well as under external climate and socio-economic stressors. Different machine learning and deep learning (DL) models have shown promising forecasting skills in various areas of application. However, their potential to forecast multi-step ahead UWD has not been fully explored. Modelling uncertain UWD patterns and accounting for variations in water demand behaviors require techniques that can extract time-varying information and multi-scale changes. In this research, we comparatively investigate different state-of-the-art machine learning- and DL-based predictive models on 1 d- and 7 d-ahead UWD forecasting, using daily demand data from the city of Milan, Italy. The contribution of this paper is two-fold. First, we compare the forecasting performance of different machine learning and DL models on single- and multi-step daily UWD forecasting. These models include an artificial neural network, a support vector regression, a light gradient boosting machine (LightGBM), and long short-term memory networks with and without an attention mechanism (LSTM and AM-LSTM). We benchmark their prediction accuracy against autoregressive time series models. Second, we investigate the potential enhancement in predictive accuracy by incorporating the wavelet transform and feature selection performed by LightGBM into these models. Results show that, overall, wavelet-enhanced feature selection improves the model predictive performance. The hybrid model combining wavelet-enhanced feature selection via LightGBM with LSTM (WT-LightGBM-(AM)-LSTM) can achieve high levels of accuracy with Nash-Sutcliffe Efficiency larger than 0.95 and Kling-Gupta Efficiency higher than 0.93 for both 1 d- and 7 d-ahead UWD forecasts. Furthermore, performance is shown to be robust under the influence of external stressors causing sudden changes in UWD
Advancing Residential Water Management by Smart Metering and Data Intensive Modeling of Consumers’ Behaviors
No full textUrban population growth, climate and land use change are boosting residential water demand. Developing suitable demand-side management strategies is essential to meet future water demands, pursue water savings, and reduce the costs for water utilities. Yet, the effectiveness of water demand management actions relies on our understanding of water consumers’ behavior. While low spatial and temporal resolution water consumption data, as traditionally gathered for billing purposes, hardly support this understanding, the advent of smart metering technologies allowed for quasi real-time monitoring water consumption at the single household level. This advanced our ability in characterizing consumers’ behavior and designing user-oriented residential water demand management strategies.
With this work, we revise consolidated practices, identify emerging trends and highlight the challenges and opportunities for future developments given by the use of smart meters. Furthermore, we present the EU-funded SmartH2O project, which aims at creating an ICT platform able to (i) capture and store quasi real time, high resolution residential water usage data, (ii) infer the main determinants of residential water demand and build customers’ behavioral models and (iii) predict how the customer behavior can be influenced by various water demand management strategies. The project exploits a social computing approach for raising users’ awareness about water consumption and pursuing water savings in the residential sector. Preliminary models and results on users behavior modelling and users segmentation using innovative fully automated algorithms are presented
Share for care. Tecnologie ed inclusione sociale per lo sviluppo dei quartieri in formali di Guayaquil, Ecuador.
No full textIl paper si inserisce nell’ambito dello studio di strategie atte al miglioramento della qualità di vita negli slum. L’elaborato presenta IMPARAR, progetto di ricerca il cui scopo è promuovere l’empowerment della popolazione dei quartieri informali della città di Guayaquil, Ecuador, caratterizzati da un limitato accesso a servizi e risorse locali. Verranno descritti gli esiti della ricerca, che hanno portato alla definizione di un sistema biunivoco di comunicazione-intervento tra le comunità locali e la Municipalità, composto da una Community Based Organization, istituzione locale che rappresenta gli abitanti stessi e i loro bisogni, e un Digital Tool System, sistema che utilizza in maniera innovativa tecnologie diffuse tra la popolazione per facilitare l’accesso a servizi e risorse. Una reale implementazione del sistema porterebbe vantaggi rilevanti sia per la pubblica amministrazione che per la comunità: un miglioramento della comunicazione e collaborazione creerebbe le condizioni necessarie per la definizione di strategie di sviluppo sostenibile
Share for Care. Communication Technologies and Social Inclusion for Empowerment in Guayaquil, Ecuador
No full textThis paper presents the main deliverable of Improving Access to Resources at Reduced Risk for Urban Areas with Strong Informal Settlements (IMPARAR), a project belonging to the realm of social innovation and development. IMPARAR aims at promoting social development and inclusion in the deprived neighborhoods of Guayaquil (Ecuador) by increasing citizen accessibility to public resources and services. Living conditions in informal settlements present several challenges, such as the lack of community management strategies, especially regarding access to services and risk prevention. Key determinants for such challenges are the poor communication between local communities and public administration, and the lack of territorial data in informal contests. To address these challenges as output of IMPARAR, we designed Share for Care, a two-way communication-intervention system, which is a composed solution characterized by a Digital Tool System and a Community-Based Organization. Share for Care is designed to support the communitarian and collaborative work of the population, by means of Information and Communication Technologies (ICTs). In the paper, the features of the system as well as the participatory approach adopted along the whole IMPARAR project are described. By combining communication technologies with policies of inclusion and social mobilization, Share for Care is designed to provide benefits to both public administrations and local communities
A multi-objective optimization framework to model 3D river and landscape evolution processes
No full textDeveloping a stochastic simulation model for the generation of residential water end-use demand time series
Full text link: Smart metering technologies allow for gathering high resolution water demand
data in the residential sector, opening up new opportunities for the development of models
describing water consumers’ behaviors. Yet, gathering such accurate water demand
data at the end-use level is limited by metering intrusiveness, costs, and privacy issues.
In this paper, we contribute a stochastic simulation model for synthetically generating
high-resolution time series of water use at the end-use level. Each water end-use fixture
in our model is characterized by its signature (i.e., its typical single-use pattern), as
well as frequency distributions of its number of uses per day, single use duration, time
of use during the day, and contribution to the total household water demand. The model
relies on statistical data from a real-world metering campaign across 9 cities in the US.
Showcasing our model outputs, we demonstrate the potential usability of this model for
characterizing the water end-use demands of different communities, as well as for analyzing
the major components of peak demand and performing scenario analysis
Unveiling Residential Water Consumers’ Behaviour and Profiles Through Machine Learning Techniques
No full textThe continuous development of urban areas worldwide in the near future is foreseen to boost household water demand, thus placing a challenge to the distribution and supply of drinking water. Whereas several studies demonstrated the potential of customized demand management strategies to pursue water saving attitudes in the residential sector, still their effects rely on the level of understanding we have about consumers’ typical behaviours. Retrieving information on users’ behaviors at the household level, as well as their explanatory and/or causal factors, is key to spot areas towards which water saving efforts can be prioritized. This, in turn, aids the design of personalized water demand management strategies, such as education campaigns and recommendations and, coupled with monitoring programs, allows evaluating their effects in terms of behavioral change and customers’ engagement.
In this work, we contribute a data-driven approach to identify and model household water users’ consumption profiles. State-of-the-art clustering methods are coupled with machine learning techniques with the aim of extracting predominant user behaviors from a set of water consumption data collected at the household scale. This allows identifying heterogeneous groups of consumers from the studied sample, as well as characterizing them with respect to several consumption features.
The approach we propose in this work is validated onto a real-world household water consumption dataset, showing its potential for understanding and modeling consumers’ profiles, as well as data mining the structure of the considered community with respect to water consumption habits, ultimately informing the bottom-up collaboration between managers and customers
Profiling residential water users’ routines by eigenbehavior modelling
Full text linkDeveloping effective demand-side management strategies is essential to meet
future residential water demands, pursue water conservation, and reduce the costs for
water utilities. The effectiveness of water demand management strategies relies on our
understanding of water consumers’ behavior and their consumption habits and routines,
which can be monitored through the deployment of smart metering technologies and
the adoption of data analytics and machine learning techniques. This work contributes
a novel modeling procedure, based on a combination of clustering and principal component
analysis, which allows performing water users’ segmentation on the basis of
their eigenbehaviors (i.e., recurrent water consumption behaviors) automatically identified
from smart metered consumption data. The approach is tested against a dataset
of smart metered water consumption data from 175 households in the municipality of
Tegna (CH). Numerical results demonstrate the potential of the method for identifying
typical profiles of water consumption, which constitute essential information to support
residential water demand management
Exploring 3D optimal channel networks by multiple organizing principles
No full textCatchment topography and flow networks are shaped by the interactions of water and sediment across various
spatial and temporal scales. The complexity of these processes hinders the development of models able to assess the
validity of general principles governing such phenomena. The theory of Optimal Channel Networks (OCNs) proved
that it is possible to generate drainage networks statistically comparable to those observed in nature by minimizing
the energy spent by the water flowing through them. So far, the OCN theory has been developed for planar 2D
domains, assuming equal energy expenditure per unit area of channel and, correspondingly, a constant slopedischarge
relationship. In this work, we apply the OCN theory to 3D problems by introducing a multi-principle
minimization starting from an artificial digital elevation model of pyramidal shape. The OCN theory assumption of
constant slope-area relationship is relaxed and embedded into a second-order principle. The modelled 3D channel
networks achieve lower total energy expenditure corresponding to 2D sub-optimal OCNs bound to specific slopearea
relationships. This is the first time we are able to explore accessible 3D OCNs starting from a general DEM.
By contrasting the modelled 3D OCNs and natural river networks, we found statistical similarities of two indexes,
namely the area exponent index and the profile concavity index. Among the wide range of alternative and suboptimal
river networks, a minimum degree of 3D network organization is found to guarantee the indexes values
within the natural range. These networks simultaneously possess topological and topographic properties of real
river networks. We found a pivotal functional link between slope-area relationship and accessible sub-optimal 2D
river network paths, which suggests that geological and climate conditions producing slope-area relationships in
natural basins co-determine the degree of optimality of accessible network paths
- …
