DMA Structured State-Estimation for Demand Monitoring

While state-estimation is widely used for electrical power systems, it is still in its infancy for water distribution systems. The main reason for this is that water systems, being underground, are much less accessible for metering. Traditionally, flow and pressure telemetry has been supplemented with pseudo-demand data in order to achieve sufficient data redundancy. Since the pseudo-demands are already estimates, there is a danger of circular conclusions. In this work, the least-squares state-estimation equations are re-formulated such that normalised demand factors within District Metered Areas (DMAs) are used instead of pseudo-measurements. This formulation has several advantages. Firstly, one source of error is eliminated since a nominal time profile for the pseudo-demands is no longer required. Secondly, the demand factors are in some sense orthogonal to the DMA boundary measurements such that swamping of the real flow measurements is avoided. The practical implementation of the above state-estimation method for large systems, however, is non-trivial. Considerable numerical analysis has been employed in order to solve large systems without shortcut approximations. Results will be presented for a 5000 node South London test area covering 88000 properties.