IEEE TRANSACTIONS ON SMART GRID, VOL. 3, NO. 2, JUNE 2012 915 Trade-Offs in PMU Deployment for State Estimation in Active Distribution Grids Junqi Liu, Graduate Student Member, IEEE, Junjie Tang, Graduate Student Member, IEEE, Ferdinanda Ponci, Senior Member, IEEE, Antonello Monti, Senior Member, IEEE, Carlo Muscas, Member, IEEE, and Paolo Attilio Pegoraro, Member, IEEE Abstract—Monitoring systems are expected to play a major role in active distribution grids, and the design of the measurement in- frastructure is a critical element for an effective operation. The use of any available and newly installed, though heterogeneous, metering device providing more accurate and real-time measure- ment data offers a new paradigm for the distribution grid moni- toring system. In this paper the authors study the meter placement problem for the measurement infrastructure of an active distri- bution network, where heterogeneous measurements provided by Phasor Measurement Units (PMUs) and other advanced measure- ment systems such as Smart Metering systems are used in addi- tion to measurements that are typical of distribution networks, in particular substation measurements and a-priori knowledge. This work aims at defining a design approach for finding the optimal measurement infrastructure for an active distribution grid. The design problem is posed in terms of a stochastic optimization with the goal of bounding the overall uncertainty of the state estimation using heterogeneous measurements while minimizing the invest- ment cost. The proposed method is also designed for computational efficiency so to cover a wide set of scenarios. Index Terms—Power system state estimation, power distribu- tion, power system measurements, uncertainty, phasor measure- ment units (PMUs). I. INTRODUCTION T HE power distribution grids are undergoing fundamental changes towards more dynamic and complex structures, due to the integration of distributed generation (DG) sources, entities with dual load-generator behavior, distributed energy storage, new equipment and services, such as intelligent elec- tronic devices (IEDs) and Smart Meters, demand side manage- ment and secondary reserve, as discussed in [1], [2]. The par- ticipation of the dual load-generator entities (prosumers) to the balancing of the distribution network changes the definition of conventional loads and requires new load modeling approach. The topology of the distribution network will undergo a transi- tion from a mainly radial to a more meshed topology with bi-di- rectional power flows. In this context, traditional passive distri- Manuscript received November 22, 2010; revised June 27, 2011, March 02, 2012; accepted March 13, 2012. Date of publication May 11, 2012; date of cur- rent version May 21, 2012. J. Liu, J. Tang, F. Ponci and A. Monti are with the Institute for Automa- tion of Complex Power Systems of the E.ON Energy Research Center at RWTH Aachen University, Aachen, Germany (e-mail: [jliu, jtang, fponci, amonti]@eonerc.rwth-aachen.de). C. Muscas and P. A. Pegoraro are with the Department of Electrical and Elec- tronic Engineering at the University of Cagliari, Cagliari, Italy (email: (carlo, paolo.pegoraro)@diee.unica.it) Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TSG.2012.2191578 bution grids are being transformed to future distribution grids with active nature. This scenario creates new challenges for the operation of the distribution networks, and advanced monitoring, control, and protection must be based on situation awareness of the system conditions. This may not be achieved with a full deployment of instrumentation as it is economically unreasonable. A key ele- ment to achieve reliable and accurate system condition aware- ness is the distribution state estimation (DSE). DSE has been widely studied with respect to the estimation accuracy and meter placement issues [3]–[7]. In these works, traditional measure- ments, such as real substation measurements of active, reactive power and voltage magnitude provided by transducers at sub- stations and pseudo-measurements with low accuracy based on historical data are used for the state estimation. Due to the increasing dynamics and uncertainly changing behavior of the actors in distribution grids, real-time moni- toring is becoming more and more important. This requires a measurement infrastructure that is currently not yet in place in distribution grids [8]. Among the measurements that would enable tight tracking and control of the distribution grids, the synchrophasors provided by Phasor Measurement Units (PMUs) are of great interest. The PMUs provide synchronized, real-time phasor measurements of voltage and current, gener- ally augmented by measurement of frequency, rate of change of frequency, power and potentially other indirect measurements. State estimation using synchrophasor measurements has also been widely proposed for transmission networks and has been proven as a promising tool to improve the performance of the state estimation (SE) [9]–[11]. However, the methods for PMU location and the use of PMU measurements for SE, taken from previous experience in trans- mission grids, are not immediately applicable to distribution grids. In first place, the availability of measurements that char- acterizes transmission grids cannot be achieved in distribution grids with much larger number of nodes through the deployment of PMU, as this would be economically unreasonable. To com- pensate for the poor availability of direct measurements, it is usual to resume to pseudo-measurements from a priori knowl- edge [12]. But, as a consequence, the uncertainty that propa- gates to the state estimation can be very large. The use of PMUs in distribution grids has been already advocated as discussed in [13]–[16]. Phasor measurements at buses close to DG have been considered in the DSE in [16]. Recently, an increasing number of smart measurement de- vices providing high accuracy and associated with advanced communication functionality are available and applied in low- 1949-3053/$31.00 © 2012 IEEE