The growing urban population and the megacities development are going to intensify urban freshwater demand in the next decades. In addition, climate change and different precipitation distribution will increase situations of water scarcity and stress. In this background, the urban context is one of the most alarming question because urban water demand is becoming more intensive and the water a rarer resource. The development of effective demand-side management strategies is essential to meet future residential water demands, pursue water savings, and reduce costs for water utilities. Yet, the effectiveness of water demand management strategies relies on our understanding of water consumers’ behavior and their consumption habits, which can be monitored through the deployment of smart metering technologies and the adoption of data analytics and machine learning techniques. This work analyzes in detail the question of urban water management and the users profiling operation, providing an overview to the entire procedure and to the different approaches present in literature. It also 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 main aim of the work is to test the potentiality of this analysis and to understand if and how it could extend knowledge of the entire system and its variables. The approach is tested on two different case studies: the first one on a synthetically generated dataset composed by 400 houses, the second on 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 constitutes essential information to support residential water demand management strategies. Further analysis, based on this started work, will be necessary to better explore and value how this method can be useful for water demand management strategies.
La crescita della popolazione e lo sviluppo di megalopoli intensificherà la richiesta urbana di acqua potabile nei prossimi decenni. Inoltre, i cambiamenti climatici e una diversa distribuzione delle precipitazioni accentueranno le situazioni di carenza e stress idrico. In questo contesto, l’ambito urbano rappresenta una delle situazioni più allarmanti in quanto la domanda diventerà sempre più intensa e l’acqua sempre più una risorsa rara. Per soddisfare le richieste, perseguire il risparmio idrico e ridurre i costi per servizi idrici è quindi essenziale lo sviluppo di strategie adeguate di gestione della domanda idrica. Tuttavia, l’efficacia di tali strategie si basa sulla comprensione del comportamento dei consumatori e sulle loro abitudini di consumo, monitorabili attraverso la diffusione di smart meters, l’adozione di migliori analisi dei dati e tecniche di apprendimento automatiche. Questo elaborato analizza in dettaglio la questione della gestione urbana e l’operazione di profilamento degli utenti, fornendo una visione d’insieme del procedimento e dei diversi approcci presenti in letteratura. Propone, poi, una nuova procedura di modellizzazione, basata su una combinazione di clusterizzazione e di analisi delle componenti principali, che consente la suddivisione degli utenti sulla base dei loro eigenbehaviors (i.e. i comportamenti ricorrenti di consumo) identificati automaticamente dai dati di consumo raccolti; lo scopo principale del lavoro è quello di testare le potenzialità di questa metodologia e di capire se e come possa migliorare la conoscenza dell’intero sistema e delle sue variabili. Il metodo è testato su due diversi casi di studio: il primo basato su dati sinteticamente generati relativi a 400 case, il secondo su un set di dati raccolti da smart meters che hanno misurato il consumo d’acqua di 175 famiglie nel comune di Tegna (CH). I risultati dimostrano le potenzialità del metodo nell’identificazione di profili tipici di consumo d’acqua. Ulteriori analisi, sulla base di questo lavoro, saranno necessarie per valutare meglio come questo metodo potrà essere utile per migliorare le strategie di gestione della domanda idrica residenziale.
Inferring and clustering residential water consumers' routines by eigenbehavior modeling
MORO, ANDREA;RIVA, LUCA
2015/2016
Abstract
The growing urban population and the megacities development are going to intensify urban freshwater demand in the next decades. In addition, climate change and different precipitation distribution will increase situations of water scarcity and stress. In this background, the urban context is one of the most alarming question because urban water demand is becoming more intensive and the water a rarer resource. The development of effective demand-side management strategies is essential to meet future residential water demands, pursue water savings, and reduce costs for water utilities. Yet, the effectiveness of water demand management strategies relies on our understanding of water consumers’ behavior and their consumption habits, which can be monitored through the deployment of smart metering technologies and the adoption of data analytics and machine learning techniques. This work analyzes in detail the question of urban water management and the users profiling operation, providing an overview to the entire procedure and to the different approaches present in literature. It also 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 main aim of the work is to test the potentiality of this analysis and to understand if and how it could extend knowledge of the entire system and its variables. The approach is tested on two different case studies: the first one on a synthetically generated dataset composed by 400 houses, the second on 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 constitutes essential information to support residential water demand management strategies. Further analysis, based on this started work, will be necessary to better explore and value how this method can be useful for water demand management strategies.File | Dimensione | Formato | |
---|---|---|---|
MoroRiva_Thesis.pdf
accessibile in internet per tutti
Descrizione: Testo della tesi
Dimensione
2.27 MB
Formato
Adobe PDF
|
2.27 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/120445