Watertown, CT, United States
Watertown, CT, United States

Time filter

Source Type

Behandish M.,University of Connecticut | Wu Z.Y.,Bentley Systems Incorporated
Procedia Engineering | Year: 2014

In spite of the growing computational power offered by the commodity hardware, fast pump scheduling of complex water distribution systems is still a challenge. In this paper, the Artificial Neural Network (ANN) meta-modeling technique has been employed with a Genetic Algorithm (GA) for simultaneously optimizing the pump operation and the tank levels at the ends of the cycle. The generalized GA+ANN algorithm has been tested on a real system in the UK. Comparing to the existing operation, the daily cost is reduced by about 10 - 15%, while the number of pump switches are kept below 4 switches-per-day. In addition, tank levels are optimized ensure a periodic behavior, which results in a predictable and stable performance over repeated cycles © 2013 The Authors.


Wu Z.Y.,Bentley Systems Incorporated | Song Y.,University of Connecticut
14th Water Distribution Systems Analysis Conference 2012, WDSA 2012 | Year: 2012

Pressure loggers are used to record field pressure data, which are the essential information for water utilities to achieve sound model calibration. Thus determining where and how many pressure loggers to be placed in a distribution system is an important task. In this paper, an effective method is presented for optimizing pressure logger placement. It is developed into a two-stage solution method. At the first stage, a Mont Carlo method is used to generate a large number of random events, each of them represents either demand change and/or leakage flow. All the events are simulated by conducting hydraulic model analysis. The simulated nodal pressures are compared with the baseline condition. The pressure change at each node is evaluated with respect of logger accuracy. If the pressure change is greater than the logger accuracy, a value of 1 is assigned to the node for an event, otherwise a 0 is a assigned to the node. Thus a binary database is constructed for the randomized event for a given distribution system. The constructed binary database is employed to optimize the pressure logger locations in the second stage. The pressure logger locations are optimized for a given number of loggers such that the randomized events are detected or covered as many as possible. The method is tested on two systems. The results are compared with the pressure logger locations previously designed by the experienced engineers. It shows the method is able to achieve greater coverage with less number of pressure loggers than the sampling design by experienced engineers. Copyright © (2012) by Engineers Australia.


Wu Z.Y.,Bentley Systems Incorporated | Behandish M.,University of Connecticut
14th Water Distribution Systems Analysis Conference 2012, WDSA 2012 | Year: 2012

This paper presents a real-time pump scheduling (RTPS) methodology and the case study of a large water distribution system. The method employs the artificial neural network (ANN) based on graphics processing unit (GPU), a meta-model or surrogate solver for evaluating the hydraulic responses of pump scheduling solutions. A genetic algorithm (GA) is used to search for nearoptimal pump operation policy (POP) subject to water supply service requirements. The resulting POP is applied to the system operation for a predefined period or so-called rolling-forward time step that is much less than a typical operation cycle (e.g. 24 hours). A rolling-forward time step is determined such that the surrogate model is capable of predicting sufficiently accurate hydraulic responses for the purpose of GA solution evaluations. At the end of each rolling forward step, new boundary conditions, e.g. tank levels obtained from the field monitoring system, are used as the initial conditions for the next GA+ANN optimization process. This RTPS methodology has been successfully applied to identify the tank operation range for real-time operation and minimize the pumping energy cost of a large demand monitoring zone (DMZ) in the UK. The results indicate that significant saving in energy costs can be achieved in comparison with the current operation policies. Copyright © (2012) by Engineers Australia.


Wu Z.Y.,Bentley Systems Incorporated | Khaliefa M.,University of Connecticut
World Environmental and Water Resources Congress 2012: Crossing Boundaries, Proceedings of the 2012 Congress | Year: 2012

Cloud computing is quickly becoming an innovative model for delivering IT infrastructure, applications and data management. It shifts the emphasis from static, stand-alone applications to dynamic, shared environments, dynamically allocated among various tasks and accessed via a network. In this paper, we investigate the use of cloud computing for high performance optimization of water distribution systems. The paper covers the general survey of leading commercial cloud computing services, high performance computing (HPC) cloud differentiators and demonstration of the improved HPC cloud implementation. With necessary background information on cloud computing, a prototype of the high performance computing (HPC) cloud is proposed and developed for water system optimization. The prototyped HPC cloud is constructed by using many-core machines that form the cloud platform for running parallel applications. Finally, as an example of cloud-based water distribution optimization, a pump scheduler has been deployed onto the HPC cloud with web-based user interface, through which a user could submit, execute and retrieve optimization analysis jobs. © ASCE 2012.


Wu Z.Y.,Bentley Systems Incorporated | Behandish M.,University of Connecticut
14th Water Distribution Systems Analysis Conference 2012, WDSA 2012 | Year: 2012

This paper presents the comparison of two different approaches for solving the computationally intensive problem of near-optimal pump scheduling for large water distribution systems. The optimization problem is formulated to minimize the pump operation cost subject to water supply service requirements. The first method, utilizes the hydraulic model integrated with a parallel genetic algorithm (GA), which can run on either a many-core machine or a cluster of many-core machines. The second method, on the other hand, uses a GPU-Accelerated artificial neural network (ANN) meta-model to surrogate the hydraulic model in GA optimization. The study shows that the GPU-based ANN is capable of rapidly predicting the energy rates with adequate accuracy and robustness, as well as the tank levels which can be used for online optimization on a rolling-forward basis. GA+ANN is capable of reducing the optimization run time from several hours to a few minutes, thus enables real-time or online pump scheduling for large water distribution systems. Copyright © (2012) by Engineers Australia.


Giustolisi O.,University of Bari | Walski T.M.,Bentley Systems Incorporated
Journal of Water Resources Planning and Management | Year: 2012

Solving water distribution network hydraulics depends to a greatextent on demand representation in the related simulation models. The classical approach of simulation models for water distribution networks (WDNs) is described as demand-driven. The demands are fixed a priori in the model as an assumption or from field observations. Recentlya more realist approach to predict the hydraulic system behavior, described as head/pressure-driven, better accounts for the fact that thedemands depend in some ways on head status of the network. Thus, thispaper presents a comprehensive view of demands in the enhanced WDN simulation models, including considerations of humanbased, volume-based,uncontrolled orifice-based, and leakage-based demands as distinct types of network outflows. The paper proposes and discusses the representation of each type of demand in a comprehensive framework that is consistent with the hydraulic principles and the specific working condition. © 2012 American Society of Civil Engineers.


Giustolisi O.,University of Bari | Berardi L.,University of Bari | Walski T.M.,Bentley Systems Incorporated
Journal of Hydroinformatics | Year: 2011

The Colebrook-White formulation of the friction factor is implicit and requires some iterations to be solved given a correct initial search value and a target accuracy. Some new explicit formulations to efficiently calculate the Colebrook-White friction factor are presented herein. The aim of this investigation is twofold: (i) to preserve the accuracy of estimates while (ii) reducing the computational burden (i.e. speed). On the one hand, the computational effectiveness is important when the intensive calculation of the friction factor (e.g. large-size water distribution networks (WDN) in optimization problems, flooding software, etc.) is required together with its derivative. On the other hand, the accuracy of the developing formula should be realistically chosen considering the remaining uncertainties surrounding the model where the friction factor is used. In the following, three strategies for friction factor mapping are proposed which were achieved by using the Evolutionary Polynomial Regression (EPR). The result is the encapsulation of some pieces of the friction factor implicit formulae within pseudo-polynomial structures. © IWA Publishing 2011.


Wu Z.Y.,Bentley Systems Incorporated | Song Y.,University of Cocnnecticut
Procedia Engineering | Year: 2014

During a hydrant flow test, the more pipes with the increased flow velocity, the better the hydrant flow test is. Therefore, a good selection of hydrants for flow test is anticipated with the flow velocity increment or extra head losses occurred in as many pipes as possible. In order to maximize the performance of flow tests, a new method is developed to search for a combination of available hydrants for flow tests so that the total length of the pipes with the increased flow velocity and/or the increased unit hydraulic head loss greater than a prescribed threshold is maximized. © 2013 The Authors. Published by Elsevier Ltd.


Syed J.L.,Al Ain Distribution Company | Wu Z.Y.,Bentley Systems Incorporated
World Environmental and Water Resources Congress 2012: Crossing Boundaries, Proceedings of the 2012 Congress | Year: 2012

Water transmission main is subjected to hydraulic transients due to sudden pipe failure, valve closure or any other transient phenomenon. These hydraulic transients might cause adverse effect within the pumps, bursting of the transmission main and detachment of the pipe from the joint locations such as air valve or control valve. The negative transient pressures can cause water quality issues especially from intrusion of contaminants from at the air valves and loose joints. In order to avoid both unacceptable minimum and maximum transient pressures, surge protection is imperative for a water transmission main. It is a common practice to locate the surge vessel near the pumps in order to dampen the generated surge due to sudden pump failure. In this study, surge analysis results considering two different topology configurations (Option 1 & 2) with respect to different surge vessel locations are compared for a water transmission main connected to two pumps (each having a capacity of 14.5 l/s @ 110 m head ) in parallel. The transmission main is serving residential consumers which are located on 50 m higher elevation from the pump station at a distance of about 12.5 KM. For Option 1, it is considered that two surge vessels are placed on a 150 mm DI line, 2.5 m far at the downstream side of each pump. For Option 2, only one surge vessel is placed on a 300 mm transmission main originating just after the parallel connection of the 150 mm DI lines at the downstream of the two pumps. Transient analysis is undertaken for different vessel sizes of two configurations. The results obtained are compared for the Minimum Transient Pressures in the modelled pipeline. Although, comparative difference between the two configurations is not much but gives a clear indication that surge vessel location and sizes could play a vital role in controlling the transient pressures. It is also observed that the pre-charge pressure in the surge vessel is also very important for controlling the transient pressures and must be precisely determined with respect to the mechanism of the working of the selected surge vessel. © 2012 ASCE.


Grant
Agency: GTR | Branch: EPSRC | Program: | Phase: Research Grant | Award Amount: 72.25K | Year: 2011

The AEC (Architecture, Engineering and Construction) industry is a highly fragmented data intensive project-based industry depending on a large number of very different professions and firms, with strong data sharing requirement across lifecycle stages from concept design to demolition. The process of designing, re-purposing, constructing and operating a building involves not only the traditional disciplines (Structure, Mechanical & Electrical, etc.) but also many new professions in areas such as energy, environment, waste, and assisted living with large data sharing requirements. In this context, data management support for the project lifecycle tends to be fragmented with a lack of an overall (project wide) data management policy. Additionally, data sets relating to a particular project can often be stored in: (i) local computers of designers/architects - often with limited network connectivity, persistence and availability; (ii) independently managed, single company-owned archives - where access is dictated by a company specific policy or by a charging model; (iii) shared archives owned by a consortium, often in the context of a particular building project - based, at best, on access policy associated with the project. The CloudBIM proposal explores the feasibility and potential for utilizing Cloud capability to address data storage and processing needs of stakeholders in the AEC (Architecture, Engineering and Construction) sector, with a view of delivering a cloud platform for research. CloudBIM will involve close consultation and interaction with major participants in the area to assess stakeholders perceptions about outsourced, virtualized Cloud storage for supporting multi-site, multi-team collaborative projects. A prototype cloud platform (based on CometCloud - www.cometcloud.org) and associated governance model will be developed and made available to the AEC research community. The project will deliver several reports based on a number of people-based activities, involving BRE (Building Research Establishment) and MBEKTN (Modern Built Environment Knowledge Transfer Network), along with a prototype using real project case studies BIM (Building Information Model) data to be provided by Bentley. A key outcome will be to spur a wide range of research-oriented activities through a strategic roadmap aimed at the exploitation of the resulting CloudBIM platform.

Loading Bentley Systems Incorporated collaborators
Loading Bentley Systems Incorporated collaborators