TCLAS - Telemetry Confidence Limit Analysis

TCLAS Rationale

Present day commercial simulation packages are well capable of efficient calculation of flows and pressures in complex water distribution systems. However, no simulation software can calculate accurate results from inaccurate data. This is the reason why the accurate (from the computational viewpoint) results of simulations can be very inaccurate when compared to the actual pressure and flow measurements. The only way to improve the accuracy of computer simulations is to use greater number of accurate measurements as input data. Unfortunately, deciding how many measurements are needed for the simulation results to achieve a required level of accuracy is a complex problem. There are many interdependent factors that need to be taken into account. These include: the number and accuracy of existing meters in the system, the location of these meters, the connectivity of network nodes and the operational state of the system.

TCLAS takes into account these interdependencies and calculates Confidence Limits on flows and pressures that are due to a given metering configuration. The results are presented graphically allowing the operator to see which parts of the network require additional meters. By exploring various metering configurations TCLAS user can find an appropriate balance between the Confidence Limits and the cost of metering. To illustrate the use of the TCLAS software a case study for a small/medium size network is included below.

The Users

The TCLAS software has been developed in the context of water distribution systems but it is of potential use to other public utilities industries such as electric power or gas distribution networks. TCLAS is a self-contained package that can be used off-line or be interfaced into an operational decision support system.

Meter Placement Design - Case Study

The sample network consists of 65-nodes and 92 pipes and it is supplied from five sources: two boreholes, one high-pressure zone inlet and two interlinked reservoirs which are replenished from an external source. All supplies to the network are measured with +-5% accuracy and a reference pressure in one of the reservoirs (top left corner on the diagram) is measured with an accuracy +-0.02m. The consumption estimates in nodes fitted with off-line meters (large consumers) were created by apportioning an appropriate fraction of a daily consumption according to a diurnal consumption pattern. These were assumed to be accurate to within +-10%. The estimates of consumptions in all other nodes were obtained by dividing the reminder of the total supply in proportion to the population count associated with a given node. Such estimates were believed to be much less accurate, with accuracies ranging from +-20% for nodes with large volume of consumptions to +-70% for nodes with small volume of consumptions.

The result of the Confidence Limit Analysis on this network is illustrated in Figure 1. The red error blocks, associated with each node, indicate that the corresponding pressures calculated by any simulation software cannot be more accurate than approx. +-1m Aq, and indeed, in some nodes it will be less accurate than +-2m Aq. This is a very significant result since the pressure difference between the highest- and the lowest-pressure node in this network is also 2m Aq. This means that the inaccurate knowledge of consumptions makes it impossible to be certain about the direction of flow in a number of pipes, let alone the actual value of these flows. The "zoom-1" display gives numerical detail of flows and pressures with their associated confidence limits in the vicinity of a selected node 22.

To investigate the effect of additional metering, a pressure meter has been added in node 22. The assumed accuracy of this meter is +-0.03m Aq which improves significantly on the previous error bound for the pressure of +-1.28m Aq. The new confidence limits calculated by TCLAS are displayed in Figure 2. It can be seen that a number of nodes, in the neighbourhood of node 22, have benefitted from the addition of this meter and now they have blue error blocks indicating that the pressure inaccuracies in these nodes are between +-0.02 to +-0.2m Aq. Additionally, a number of nodes that lie further away from the node 22 had their error bounds reduced to below +-1.0m Aq (partially filled red blocks). It is interesting to note however, from the "zoom-2" display, that the addition of a single pressure meter did not significantly improve the confidence limits on flows. Consequently, if the flows need to be known with greater confidence, additional meters are needed.

The design of a telemetry system proceeds as an iterative process aiming at achieving acceptable error bounds with a minimum number of meters. Such an iterative design has a distinct advantage that the design engineer gains an in-depth knowledge of the effects of individual meters thus assisting in planning of various contingencies.

TCLAS - Screen Shots

Figure 1: Confidence limits for a network with one reference pressure measurement only

Zoom-page for the selected node 22

Figure 2: Confidence limits for a network with one additional pressure measurenent in node 22

Zoom-page for the node 22 (with an additional pressure measurement in node 22)

Publications

• Bargiela A., Hainsworth G., Pressure and flow uncertainty in water systems, ASCE Journal of Water Resources Planning and Management, Vol 115, 2, March 1989, pp. 212-229, PDF
• Bargiela A., Hainsworth G., Pressure and flow uncertainty in water systems, Water Resources Journal, Sept., 1990, pp. 40-48. This paper was selected for a reprint from the ASCE Journal, Vol.115, 2
• Hainsworth G., Bargiela A., Optimal telemetry system for water networks, Int. Symposium on Optimal Modeling of Water Distribution Systems, Lexington, K.Y., May 1988.
• Bargiela A., A graphical interactive software tool for confidence limit analysis, AWWA Conference on Computers and Automation in the Water Industry, Denver, CO, 2-4 April 1989.
• Bargiela A., Hainsworth G., Confidence limit analysis in water systems, pp. 43-58, in Computer Applications in Water Supply, Vol 2. Systems Optimisation, Coulbeck B., Orr C.H. (Eds.), Research Studies Press U.K., John Wiley and Sons, 1988, ISBN: 0 471917842, PDF
• Bargiela A., Hainsworth G., Telemetry design with TCLAS software, pp. 1-11, in Computer Applications in Water Supply, Vol 2. Systems Optimisation, Coulbeck B., Orr C.H. (Eds.), Research Studies Press U.K., John Wiley and Sons, 1988, ISBN: 0 471917842, PDF
• Bargiela A., Managing uncertainty in operational control of water distribution systems", in Integrated Computer Applications, Vol 2, (Ed.) B Coulbeck, J Wiley, 1993, ISBN: 0 471 94232 2, pp 353-363.
• Hartley J., Bargiela A., TPML: Parallel meta-language for scientific and engineering computations using transputersProceedings of 2nd Int. Conference on Software for Supercomputers and Multiprocessors. SMS-94, Moscow 1994, PDF
• Hartley J., Bargiela A., XTPML - Simplifying the Development of Parallel Programs for Implementation on Various Transputer Architectures, Proc. European Simulation Symposium ESS-98, Oct. 1998, ISBN 1-56555-147-8, pp.119-123, PDF
• Hartley J., Bargiela A., Probabilistic Simulation of Large-scale Water Distribution Systems, Proceedings of European Simulation Symposium ES-96, Genoa, October 1996, ISBN 1-565555-099-4 (Vol.2), pp.403-407, PDF
• Hartley J., Bargiela A., Cant R., Parallel simulation of large scale water distribution systems, Proceedings of Modelling and Simulation Conference ESM-95, Prague, June 1995, ISBN 1-56555-080-3, pp. 723-727, PDF
• Hartley J., Bargiela A., Parallel State Estimation with Confidence Limit Analysis, Parallel Algorithms and Applications , Vol. 11, No.1-2, 1997, pp. 155-167, PDF
• Bargiela A., Nonlinear network tearing algorithm for transputer system implementation, Proc. of Int. Conf. TAPA-92, Melbourne, November 1992, ISBN 905199115 0, pp. 19-24
• Bargiela A., Hosseinzaman A., Parallel simulation of nonlinear networks using diakoptics, in Parallel Computing and Transputer Applications, M. Valero (ed.), IOS Press/CIMNE, Barcelona, 1992, ISBN 84 87867 138, pp. 1463-1473.