Optimal sensor placement for classifier-based leak localization in drinking water networks

Soldevila, Adrià; Tornil-Sin, Sebastiàn; Fernandez-Canti, Rosa M.; Blesa, Joaquim; Puig, Vicenç

doi:10.1109/systol.2016.7739771

Cited by 3 publications

(5 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This might be due to the Kriging, in which the average error in this case was the worst compared to all the other sensor placement. This is not a new result, as it is known from posterior works that the placement of the sensors will produce very different results [39]. For the validation data set, the evolution over time of the Bayes time reasoning is plotted in Figure 8, which provides an accuracy boost in less than 5 h with a significant increase.…”

Section: Resultsmentioning

confidence: 79%

Leak Localization in Water Distribution Networks Using Pressure and Data-Driven Classifier Approach

Sun

Parellada

Puig

et al. 2019

Water

View full text Add to dashboard Cite

Leaks in water distribution networks (WDNs) are one of the main reasons for water loss during fluid transportation. Considering the worldwide problem of water scarcity, added to the challenges that a growing population brings, minimizing water losses through leak detection and localization, timely and efficiently using advanced techniques is an urgent humanitarian need. There are numerous methods being used to localize water leaks in WDNs through constructing hydraulic models or analyzing flow/pressure deviations between the observed data and the estimated values. However, from the application perspective, it is very practical to implement an approach which does not rely too much on measurements and complex models with reasonable computation demand. Under this context, this paper presents a novel method for leak localization which uses a data-driven approach based on limit pressure measurements in WDNs with two stages included: (1) Two different machine learning classifiers based on linear discriminant analysis (LDA) and neural networks (NNET) are developed to determine the probabilities of each node having a leak inside a WDN; (2) Bayesian temporal reasoning is applied afterwards to rescale the probabilities of each possible leak location at each time step after a leak is detected, with the aim of improving the localization accuracy. As an initial illustration, the hypothetical benchmark Hanoi district metered area (DMA) is used as the case study to test the performance of the proposed approach. Using the fitting accuracy and average topological distance (ATD) as performance indicators, the preliminary results reaches more than 80% accuracy in the best cases.

show abstract

Section: Resultsmentioning

confidence: 79%

Leak Localization in Water Distribution Networks Using Pressure and Data-Driven Classifier Approach

Sun

Parellada

Puig

et al. 2019

Water

View full text Add to dashboard Cite

show abstract

“…2. A previous work [10] states that the optimal number of sensors for this network are two by considering the relationship between installation costs, maintenance, and the information obtained from the network. Therefore, as a reference point, we also consider two sensors to be placed in the network.…”

Section: Resultsmentioning

confidence: 99%

“…, q N ) are not measured. In real WDN, the estimated demands are typically calculated based on the total demand of the network as a distribution at the nodal level according to historical consumption records [10]. This way of estimating the demand is also considered for some academic benchmark such as the Hanoi network.…”

Section: Background On Leak Localizationmentioning

confidence: 99%

“…In the literature, there are different methods for sensor placement based on leak location error minimization. Some of them consider metaheuristic such as genetic algorithms [9], [10] or particle swarm optimization [11]. Other works are focused on minimizing the error in the leak detection based on the information theory [4], evolutionary algorithms [12], decision-making techniques, such as DEMATEL combined with fuzzy logic [13], among others.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Pressure Sensor Placement for Leak Localization Using Simulated Annealing with Hyperparameter Optimization

Morales-Gonzalez

Santos‐Ruiz

López‐Estrada

et al. 2021

2021 5th International Conference on Control and Fault-Tolerant Systems (SysTol)

View full text Add to dashboard Cite

This paper presents a machine learning method for optimal pressure sensor placement in water distribution networks. The proposed approach considers annealing metaheuristics, and it is focused on the optimal placement of the sensors to perform leak localization. Unlike other works, this method considers a limited number of sensors to be placed and some restrictions on critical nodes that can be excluded or preselected. The approach is based on minimizing a cost function; this cost function is assigned as the leak location error, which varies depending on the subset of nodes where the sensors are assigned and the configuration of the leak location method. A leak localization technique based on k-NN classifiers was used, and during the minimization of the cost function, classifier hyperparameters were simultaneously optimized. The proposed method was tested on the Hanoi water distribution network programmed in MATLAB.

show abstract

“…conjugate gradients, Nelder-Mead, etc.) on (10). Once the parameters and hyperparameters of the GPR model are known, it is possible to make predictions with it, this means estimating the response y new for each new input x new .…”

Section: B Gaussian Process Regressionmentioning

confidence: 99%

Estimation of Node Pressures in Water Distribution Networks by Gaussian Process Regression

Santos‐Ruiz

López‐Estrada

Puig

et al. 2019

2019 4th Conference on Control and Fault Tolerant Systems (SysTol)

Self Cite

View full text Add to dashboard Cite

This paper proposes a method to predict pressures in all nodes of a water distribution network (WDN) by Gaussian process regression (GPR) from pressure measurements in a subset of selected nodes. The pressure sensors are placed in the nodes where, together, they capture the maximum pressure variance and also have a minimum sensitivity to measurement noise. As a case study, the proposed method was tested on a dataset obtained from simulations with the hydraulic model of the Hanoi WDN. Using only three pressure sensors, the GPR estimation error in the pressures of the unmeasured nodes are comparable to the error due to measurement noise in physical pressure sensors.

show abstract

Optimal sensor placement for classifier-based leak localization in drinking water networks

Cited by 3 publications

References 13 publications

Leak Localization in Water Distribution Networks Using Pressure and Data-Driven Classifier Approach

Leak Localization in Water Distribution Networks Using Pressure and Data-Driven Classifier Approach

Pressure Sensor Placement for Leak Localization Using Simulated Annealing with Hyperparameter Optimization

Estimation of Node Pressures in Water Distribution Networks by Gaussian Process Regression

Contact Info

Product

Resources

About