RESEARCH
Embedded Files
DERs Integration
DERs Integration
The traditional power system was designed to generate electricity at centralized power plants and deliver it to consumers through transmission and distribution networks. As a result, only the transmission system contained power resources capable of participating in the power market.
Recently, however, the proliferation of small-scale Distributed Energy Resources (DERs) within the distribution system has introduced new opportunities for market participation. These distributed resources, which are steadily increasing, also bring challenges related to their volatility and uncertainty. A significant issue is the lack of a dedicated entity to monitor and address physical problems caused by these DERs, such as overvoltage and capacity overloads.
To address these technical challenges, the establishment of a Distribution System Operator (DSO) has become essential for ensuring the reliable and efficient operation of distribution systems.
Machine Learning for Power Distribution Systems
Machine Learning for Power Distribution Systems
Traditionally, meters were used only for billing purposes. However, with the advent of Advanced Metering Infrastructure (AMI), vast amounts of data are now being collected. This data, when combined with advanced machine learning techniques, can significantly reduce modeling errors in the power distribution system.
AMI data serves a variety of purposes, including:
Phase Identification: Determining the phase to which each meter is connected.
Transformer - Meter Pairing Identification: Identifying the connections between distribution transformers and customers.
Parameter Estimation: Estimating the type and length of customer service lines.
Topology Identification: Mapping the configuration of the distribution system from feeders to end-users.
These applications demonstrate the transformative potential of AMI in enhancing the accuracy and efficiency of distribution system management.
DR Baseline Estimation
DR Baseline Estimation
Missing data is a common issue in load profile processing for distribution systems, often caused by temporary communication losses. Analysis of real-world smart meter data (e.g., from Pecan Street and utilities like Duke Energy) reveals that ~70% of missing data segments are under 4 hours, making their restoration critical for enhancing data quality. Improved data quality supports downstream tasks like load forecasting, non-intrusive load monitoring, and load pattern analysis.
Additionally, demand response (DR) baseline estimation can be treated as a special case of missing data restoration. For instance, during Conservation Voltage Reduction (CVR) events, utilities reduce substation voltage to lower peak loads. Estimating the "would have been" load baseline during such events involves restoring missing data for the CVR period, based on pre- and post-event load profiles. Accurate baseline estimation is vital for evaluating DR performance and benefits load service providers by enhancing system analysis and planning.
Page updated
Google Sites
Report abuse