Concept of an Integrated ICT System for the Management of the Large-scale Water Distribution Network of the Upper Silesian Waterworks in Poland

Transcription

1 Concept of an Integrated ICT System for the Management of the Large-scale Water Distribution Network of the Upper Silesian Waterworks in Poland J. Studziński*, P. Wójtowicz** and I. Zimoch*** * Polish Academy of Science Systems Research Institute, Newelska 6, Warsaw, Poland ( jan.studzinski@ibspan.waw.pl) ** Institute of Environmental Protection Engineering, Faculty of Environmental Engineering, Wroclaw University of Technology, Wybrzeże S. Wyspiańskiego 27, Wrocław, Poland ( patryk.wojtowicz@pwr.edu.pl) (corresponding author) *** Silesian University of Technology, Institute of Water and Wastewater Engineering, Konarskiego 18, Gliwice, Poland ( izabela.zimoch@polsl.pl) Abstract This paper presents a description and the planned application of an integrated ICT system and its numerical algorithms to support the complex management of a large water transfer and distribution system - the Upper Silesian Waterworks PLC (USW). The USW is the largest water company in Poland and one of the biggest in Europe. The algorithms are included as software modules within the structure of the ICT system. Their tasks are mathematical modeling, forecasting, approximation as well as optimization and control. Used in different combinations, these tasks can improve the operation and maintenance of a large municipal drinking water supply system making it more efficient, responsive and costeffective. Keywords Waterworks management, mathematical modeling, computer simulation, optimization, approximation algorithm, automatic calibration, hydraulic model, case study. INTRODUCTION The worldwide trend in the computerization of waterworks involves the implementation of integrated information communication technology (ICT) for comprehensive management and decision support (Sempere-Payá et al., 2013). An integrated management system for a water network usually consists of GIS (Geographical Information System), SCADA (System of Control and Diagnostics Analysis) and CIS (Customer Information System) systems, which are integrated strictly with some modeling, optimization and approximation algorithms. Due to this strict cooperation under several programs, all water network management tasks can be automatically executed or computer supported. These tasks concern technical as well as organizational, administrative and economical problems (Kotowski et al., 2009; Studzinski, 2013). Under these latter problems, planning the revitalization of an aging water network is of special importance as it has a key impact on the reduction of the failure rate in the water network. The aim of the current project is to design, develop and implement structures, algorithms and computer programs into a single integrated system supporting comprehensive control and management of a complex water supply system for large agglomerations of more than 1 million residents.

2 SILESIAN WATERWORKS PLC The ICT system will be implemented for the water transfer and distribution system of the Upper Silesian Waterworks PLC (USW). The USW is the largest water company in Poland and one of the biggest in Europe delivering water to over 3 million customers. It covers an area of about 4.3 thousand square kilometers and the main water network is over 880 km long. This water distribution system incorporates the central and western sub-regions of Silesia (Silesian metropolis) supplying water totaling about m 3 /d (in 2012). The USW distribution system is a mix of pressure and gravity mains (with diameters from 300 mm up to 1800 mm) conveying water to several local (municipal) distribution systems. The USW system is composed of 11 water treatment plants, 19 pumping stations and 9 complexes of storage tanks and reservoirs (with total capacity of m 3 ). The large area of the water supply system is characterized by a great diversity of terrain altitude, heavy concentration of industry (including coal mines, steel, energy, automotive, machinery and chemical) and by mining damages. The telemetric system of the USW is composed of over 150 monitoring points where pressure and flow rates are measured. In selected locations additional measurements of turbidity and chlorine residual in water are taken. In total over 500 sensors are used. The data from the SCADA system and also mobile (temporary) measurement points will be used for the calibration and validation of the network hydraulic model (Kapelan et. al., 2007; Savic et al., 2009). The problems associated with the management of the USW lie in its size, the large number of sources with varying levels of water quality, the large number of wholesale customers (coal mines) determining the flows of water in the water supply, a large diversity in the age and material of pipes, and the significant dispersion of critical facilities throughout the entire water supply system. The proposed ICT system for the management and control of the large water supply system will have an impact on the progressive elimination of the weaknesses of the USW which include an oversized production capacity in relation to the current water demand, the need for maintenance and repair of approximately 60% of the pipelines, the periodic change in the water quality and limited flexibility in market activities. Furthermore, almost half of the area covered by the USW system is under the strong influence of mining damages. The way to overcome these weaknesses is to keep the infrastructure facilities and the water treatment plants in good working order. This can be achieved by optimizing the structure of sources and the process of water distribution as well as by maintaining the distribution network in proper technical conditions through the optimal management of emergency repairs, renovations and upgrades.

3 Figure 1 Structure of the water distribution system for the Upper Silesian Waterworks. CONCEPT OF AN ICT SYSTEM FOR THE UPPER SILESIAN WATERWORKS PLC At the Systems Research Institute in Warsaw an integrated ICT system for comprehensive water network management has been developed and its structure is shown in Figure 2 (Studziński, 2013). Rehabilitation WDN optimisation Control Leakage detection Approximation Planning Optimisation algorithms Hydraulic model Water consumption forecasting CIS Data base GIS Telemetry (SCADA) Figure 2 Block diagram of the ICT system for water network management.

4 The developed ICT system consists of 22 programs cooperating with each other in different combinations depending on the tasks to be solved. Three modules are responsible for the realization of the management tasks. The management tasks are solved using the hydraulic model and optimization algorithms (in the MOSUW module), by using the approximation algorithms (in the Kriging applications module) and by using the algorithms of mathematical modeling (in the Objects identification module). All programs of the MOSUW module work with the hydraulic model of the water network and use a multi-criteria optimization algorithm for realizing the tasks concerning model calibration, water system optimization and planning, pump control and the planning of the SCADA (Straubel and Holznagel, 1999). For the solution of other tasks only multiple simulations of the hydraulic model under different working conditions of the water network are executed. Programs in the Kriging application module use the algorithms of kriging approximation that produce value distributions in graphical form of parameters connected with the water network and with its operation (Bogdan and Studzinski, 2006). In last module, Objects identification, several programs are collected to mathematically model the dynamical processes by means of time series methods including least squares algorithms such as Kalman s, Clarke s, maximum likelihood, linear and nonlinear regression algorithms (Hryniewicz and Studzinski, 2002). The key component of the ICT system is the Branch Data Base (BDB) of the GIS system that records all information and technical, technological and economical data of the water network, its objects and of the water consumers accessed to the water network. The BDB is the main source of input data for the hydraulic model that supports the calculations of all programs collected in the modules MOSUW and Kriging applications. All programs of the ICT system communicate with each other and with the Branch Data Base. These exchange files are used e.g. to generate the water network graphs from GIS to the hydraulic model (MOSUW-H) or to export the results of the hydraulic calculations to the optimization program (MOSUW-O) while optimizing the water netwok. Through this cooperation of several programs solving different management tasks for the water network, a synergy effect arises improving the efficiency of the running programs. The following are some functions of the ICT system that can be executed only with the cooperation of several functioning programs: hydraulic calculations of the water network (cooperating programs GIS, CIS and MOSUW-H for hydraulic calculations), automatic calibration of the hydraulic model (programs GIS, CIS, SCADA and MOSUW-K for calibration), discovering and localization of leakage points in the water network (GIS, SCADA, CIS and MOSUW-A for water leak detection), optimization and planning of the water network (GIS, CIS and MOSUW-O for network optimization), control of the pump sets installed in the water network (GIS, CIS and MOSUW-P for pumps control), optimal planning of the SCADA system for the water network (GIS, CIS and MOSUW-M for measurement point localization). The solving of these tasks demonstrates the importance of integrating different programs within the framework of a united information system. The appropriate programs from the MOSUW module realize the above functions; however, they have to be supplied with the input data by GIS, CIS and also by the

5 SCADA systems. The components GIS, SCADA and CIS are adopted from commercially available software and integrated within the comprehensive ICT system environment. All programs of the MOSUW module use the data computed by the hydraulic model and its calibration and validation determines the exactness and quality of all management tasks executed by the ICT system. The quality of the final hydraulic model will be assessed according to the WRC standards (for measurements of pressure and flow rate). The following paragraphs characterize selected algorithms used in the planned ICT system. ALGORITHM FOR WATER AGE CALCULATION As water age increases in the network water quality declines, which is especially the case in older water networks. In calculating the water age, the hourly curves of water consumption in the network end user nodes and the formulas for the mixing of water in the nodes and pipes must be given. Water age values are usually established after a couple of days of the given simulation time. For water mixing, the weighted average formula with (1) equal or (2) linear or (3) exponential growing weight coefficients w i is used: (1) w i = 1,0 (2) w i = 1,0 + 0,050*t (3) w i = exp(+0,030*t) with t time in days. The type of weight coefficient is to be selected arbitrarily by the program user based on knowledge about the age and material structure of the water network and the features of the water produced. At the start of the program the water age in all nodes and pipes of the network is equal to zero. For the successive time steps the hydraulic calculations are made and the water age for successive water packets moving in the pipes depending on the water consumption in the end nodes is calculated. The mixing of water occurs in the nodes and there the water changes its age while in the pipes the water packets are only pushed onwards. As the flow velocities in each pipe and the pipe lengths are known, it is possible to exactly calculate the movements of the water packets in all pipes for each time step. As a result, water packets with water of different ages can be located in the pipes. The results of the water age calculation for an exemplary water network are shown in Figure 3. On the network graph the ages of water are represented by different colors. Also included is a map of water age distribution in the network designed by means of kriging approximation (Studziński and Bogdan 2007).

6 a) b) Figure 3. Calculation results of water age (a) and its distribution (b) in the water network designed by kriging approximation. ALGORITHM FOR PLANNING WATER NETWORK REVITALIZATION Carrying out the revitalization of the water network involves the replacement of some network pipes due to their flawed technical state. In this process, new pipes of the same diameter replace older pipes. The goal of this approach is to reduce the occurrence of network breakdowns thereby limiting potential water losses in the network. Breakdowns and accidents in water networks can be a serious problem with

7 water losses reaching up to 30% of water production in older municipal waterworks leading to significant financial losses for the enterprise (Saegrov 2000). While planning the revitalization, the particular pipes to be replaced within the network are selected. The selection is done with the goal of minimizing the water network s susceptibility to accidents and to ensure proper functioning over the whole network. During the selection process the following factors are taken into consideration: Technical state of the pipes characterized by their roughness. Current durability of the pipes calculated as the difference between the year of pipe construction and the normative pipe durability. Pipe liability to break down in percent defined on the basis of historical data concerning pipe damage. Risk of water losses calculated as the pressure in the pipe modified by the pipe diameter: p (1 + d / 500). Costs of the pipe revitalization consisting of two components, i.e. the costs of the pipe installation and the cost of purchasing the new pipe. In order to select the pipes for revitalization from those within the entire network, the revitalization indicator is calculated from the following formula: IR = w c Cn + w t (1.0 - Tn) + w a An + w s Sn where w c, w t, w a and w s are weight coefficients, Cn represents pipe roughness, Tn means current pipe durability, An is pipe liability to breakdowns and Sn is the risk of water losses defined for the pipe concerned. The program user chooses the weight coefficients arbitrarily and all factors in the formula are normalized for values from 0.0 up to 1.0. After the indicators are calculated for all pipes, a ranking list for them can be prepared according to the shrinking indicator values. The costs of the revitalization of the sorted pipes can be defined considering the two cost components. Depending on the available funding, one can decide which pipes should be replaced taking the pipes from the top of the ranking and summarizing the costs of their revitalization up to the funding limit. After the pipes due for replacement are selected, the effects of the planned revitalization can be verified by performing the hydraulic calculation for the whole water network with stated roughness values equal to null for the selected pipes. The algorithm can predict the reduced vulnerability of the water network to accidents and the possible increased water pressures for some end user nodes if the revitalization action would be performed. Some exemplary results of revitalization planning are shown in Figure 4 where the pipes selected for the exchange are marked in green.

8 a) b) Figure 4. Water network revitalization; water network graph before (a) and after (b) revitalization planning with the pipes indicated for replacement (in green). NETWORK VALVE CONTROL The goal of valve control in the water network is to change the flow distribution in the network in such the way that the water velocity in the pipes increases. As a result, the cases of low velocity flow or stationary water in the network will be avoided and the quality of the water will improve. This task is solved using a heuristic multi criteria algorithm (Straubel and Holznagel 1999). At the end of the calculation a set of Paretooptimal solutions is received from which the program user shall select the one

9 solution which is best from user point of view. For the optimization two goal functions are defined as follows: 1. Maximum of minimal value of water velocity in the water network pipes. 2. Minimum of maximal deviation from the given pressure value in the water network nodes. The first goal function is the main criterion of optimization and the second function is regarded more as a restriction. Before starting the optimization run the valves to be controlled have to be provided and their characteristics of closing must be defined. Exemplary results of valve control are shown in Figure 5. a) b)

10 c) d) Figure 5. Gates control; a) water network investigated, b) setting the valves for control, c) calculating the goal criteria d) calculation results. CONCLUSIONS The use of mathematical models and optimization algorithms can help to solve problems concerning the management of complex water networks. These methods built into user-friendly integrated ICT system could became a standard information and decision support tool for management of urban water supply systems. The implementation of the ICT system in Upper Silesia Waterworks main water distribution system will help cut down the operational costs of the water network, boost its reliability and ensure a high and consistent quality of the water produced.

11 The integrated ICT system shall also improve and make easier the job of water network operators and planners and also of the management staff of the waterworks. ACKNOWLEDGMENTS The research presented in this paper has been realized in the framework of the research project of the Polish National Centre for Research and Development (NCBiR) co-financed by the European Union from the European Regional Development Fund, Sub-measure "Development Projects"; project title: IT system supporting the optimization and planning of production and distribution of water intended for human consumption in the sub-region of the central and western province of Silesia ; project ref. no. POIG /12. REFERENCES Bogdan, L. and Studziński, J. (2006) Kriging approximation: algorithms, program and calculation results. In: Studziński J., Hryniewicz O. (Eds.) Eco-Info and Systems Research. PAS SRI, Series Systems Research, 52, Warsaw. Eliades, D.G. and Polycarpou, M.M. (2012) Leakage fault detection in district metered areas of water distribution systems. J. Hydroinformatics, 14 (4), Hryniewicz, O. and Studziński, J. (2002) Application of systems analysis tools in environmental engineering. In: Gnauck A., Heinrich R. (Eds.) Information Society and Enlargement of the European Union. 17 th International Conference on Informatics and Environmental Protection, 58-68, Cottbus Kapelan, Z., Savic, D. A. and Walters, G. A. (2007) Calibration of water distribution hydraulic models using a Bayesian-Type procedure. ASCE J. Hydraul. Eng. 133 (8), Mendez, M., Araya, A. J. and Sanchez, L.D. (2013) Automated parameter optimization of a water distribution system. J. Hydroinformatics, 15 (1), Saegrov, S. (2005) Care-W Computer Aided Rehabilitation for Water Networks, IWA Publishing, Alliance House, London. Savic, D. A., Kapelan, Z. and Jonkergouw, P. (2009) Quo vadis water distribution model calibration? Urban Water J. 6 (1), Sempere-Payá, V., Todolí-Ferrandis, D. and Santonja-Climent, S. (2013) ICT as an Enabler to Smart Water Management, In: Smart Sensors for Real-Time Water Quality Monitoring, Eds. Mukhopadhyay, S. C., Mason, A., Straubel, R. and Holznagel, B. (1999) Mehrkriteriale Optimierung fuer Planung und Steuerung von Trink- und Abwasser-Verbundsystemen. Wasser Abwasser, 140 (3), (in German). Studziński, J. (2013) IT system for computer aided management of communal water networks by means of GIS, SCADA, mathematical models and optimization algorithms. In: Proceedings of the ICT for Sustainability Conference: ICT4S, Zurich, February 14-16, 2013, Kotowski A., Pawlak A., Wójtowicz P. (2010) Modelling of municipal water distribution system: A case study. Environmental Pollution Control, 32 (2), (in Polish).