Nowadays the number of active elements connected to the distribution network, such as DG, energy storage, EVs or self-producing consumers, is increasing exponentially. Therefore, the use of monitoring and diagnosis tools is crucial in distribution networks. Until a few years ago, the possibility of implementing a state estimator (SE) in LV or MV networks was not considered technically or economically feasible. The deployment of smart grids, in particular the installation of smart-meters and remote controls associated with distribution automation systems, is providing the distribution network with new information sources.
The main objective of MONICA (Spanish acronym of Advanced Monitoring and Control) project is to develop suitable monitoring and diagnostic tools specifically designed for distribution networks (MV and LV), similar to those that have traditionally existed at higher voltage levels. The core of the monitoring process included in MONICA is a SE tool. Even though the underlying philosophy and objectives are the same as that of transmission-level SE, an ad hoc development has been undertaken owing to the following distinguishing features of MV and LV grids:
Taking those peculiarities into account, the SE developed for Smartcity comprises two stages:
In an environment in which the number of measurements available increases (smart-meters at LV and MV consumers, sensors at primary and secondary substations), it is essential to use tools that take advantage of all this information. The State Estimation algorithm allows to obtain optimally the state of operation of the network using all the information supplied by the measuring instruments installed in the network. The following table shows the main differences between the SE and the traditional load flow method.
|State Estimation||Load Flow|
|Number of measurements||More measurements than states (redundancy)||Same number of measurements than states|
|Type and distribution of measurements||Greater diversity||Restricted|
|Characterization of measurements uncertainty||Possible||Impossible|
|Detection of errors in model or measurements||Possible||Impossible|
The initial scope of the project covers two MV feeders, with a total of 57 secondary substations and the associated capillary system of LV feeders delivering electricity to over 5,000 customers, each equipped with a smart meter. As described earlier, a whole network of sensors, both at MV and LV levels, has been deployed in the Smartcity project. Figure 2 schematically shows the location and type of sensors for a generic MV feeder with several secondary substations and LV customers. For convenience, the MV/LV transformer is included in the MV SE model.
At the MV level the following measurements are available every 5 minutes:
For the LV level, with a fifteen-minute time lapse, the available measurements can be summarized as follows:
The SE has been satisfactorily tested with both simulated and actual data. It provides the maximum likelihood voltage magnitudes at all buses, the currents and power flows across all feeder sections and the resulting losses. In addition, for LV feeders it allows abnormal or unacceptable system imbalances to be detected. Just as a sample, Figure 3 shows, for each hour on the 20th of August, 2016, the resulting ohmic losses for the set of three phases and the neutral cables fed from the secondary substation No. 307 in Smartcity Malaga. As can be seen, the resulting imbalance is not negligible, which has prompted the utility to undertake corrective actions aimed at more optimally relocating individual customers to achieve a better balance.
Based on the results provided by the SE, other diagnostic and corrective tools can be developed, such as:
One of the most interesting applications of the SE is the possibility of detecting wrong measurements. The following test was carried out with real data from the LV secondary substation 307, Smartcity Malaga:
The SE is a powerful tool that can be used with the following functionalities: