A New Monitoring Approach for Power Distribution Systems
Our monitoring solution for distribution systems is based on a scalable and data-driven approach. The distribution system is divided into a number of sections and the estimations are performed for each section separately. The monitoring focuses on estimation of the voltage magnitude rather than the complete system state, which is of the relevant quantity for most application. Using a data-driven technique such as artificial neural networks (ANNs), it is possible to perform estimations with no need for system model in real-time operation and with very few measurements at both MV and LV levels.
The concept
This scalable and low-cost solution can unlock the door to the development of monitoring and automation in distribution systems by overcoming the cost barriers associated with the implementation of classical approaches.
Rather than targeting a very detailed and accurate picture (i.e. through estimating the state of the system), our approach enables the operator to have a rough understanding of the ongoing grid conditions. More specifically, it enables the operator, above all, to determine if there is a voltage violation at all, and in the second instance, to detect the weakest points of the network without heavy investments.
In this respect, our solution can constitute an intermediate but affordable step in the long-term process of automation of distribution systems. Estimation of the voltage phase angles is intentionally excluded from the monitoring problem as it is currently of little practical importance for DSOs in most cases.
The information provided by such a monitoring system is beneficial both for the utility and the consumers: the utility can use the information not only to better identify operational limits of the network and take necessary corrective actions, but also to make better decisions about network reinforcement. All this is expected to improve power quality for consumers.
Key Features in a Nutshell
- Very low measurement requirements: Depending on the number and accuracy of the available measurements, the estimation accuracy will improve, but only the voltage and current at the substation are necessary in all cases.
- Focus on voltage magnitude monitoring rather the state estimation: Estimation of phase angle is excluded intentionally to simplify the problem.
- Flexible and scalable architecture: The overall monitoring system is composed of a number of estimation modules (i.e. local estimators as shown in the figure, each responsible for one section) which work independently; the overall architecture can be also extended by including more modules.
- Completely data-driven: DSOs do not need to provide grid data (topology and parameter) for tuning the estimators for their grids at the deployment phase; furthermore, grid parameters are not needed for operation in real-time.
- Robustness again measurement noise: The ANNs used in estimators are trained so that they can maintain reasonable performance in presence of measurement noise.
To ensure reliable estimation results in MV grids, where reconfiguration is common, MV local estimators are equipped with a Configuration Identification (CI) unit which allows them to identify the actual grid configuration using their measurement inputs.
Development status
Currently, a prototype of the estimator is built in the laboratory of the ACS institute. For data collection and visualization, a web-based ad hoc solution is developed. This visualization is useful not only for development and field test, but also for maintenance, as it is independent of any existing visualization system. In parallel to the described web-based visualization, an interface could be implemented to transmit data to a SCADA system using protocols like IEC 61850 or DNP3. The connection to a SCADA system would be also available locally and globally from the cloud service.
Gridhound UG (haftungsbeschränkt) Startup
In January 2015, Gridhound UG (haftungsbeschränkt) was founded as a spin-off of the ACS Institute to develop innovative grid monitoring and control services for electric power system, with a focus on distribution systems.
The main idea is to use the data-driven approach in a cloud-based environment and deliver Monitoring as a Service (Maas) for DSOs. Gridhound UG has been already selected and granted financial support by the FINODEX (Future INternet Open Data EXpansion), which is one of the 16 Future Internet Accelerator projects co-funded by the European Union. The aim of this program is to support SMEs and Web Entrepreneurs to develop products, services building upon FIWARE technologies and using Open Data.
REFERENCES
[1] M. Ferdowsi, A. Benigni, A. Löwen, B. Zargar, A. Monti, and F. Ponci, “A Scalable Data-driven Monitoring Approach for Distribution Systems,” IEEE Transaction on Instrumentation and Measurements, Special Issue of I2MTC 2014, Early access available online.
[2] M. Ferdowsi, A. Benigni, A. Löwen, P. McKeever, A. Monti, and F. Ponci, “New Monitoring Approach for Distribution Systems,” 2014 IEEE International Instrumentation and Measurement Technology Conference (I2MTC), 12-15 May 2014.
[3] M. Ferdowsi, B. Zargar, F. Ponci and A. Monti, “Design Considerations for Artificial Neural Network-based Estimators in Monitoring of Distribution Systems,” 2014 IEEE International Workshop on Applied Measurements for Power Systems (AMPS 2014), 24-27 September 2014, Aachen, Germany.