Smart Meter Phase Identification Using Constrained Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for phase identification in electric power distribution systems are inaccurate due to phasing errors, computational complexity, and high costs associated with equipment installation, especially when dealing with distributed energy resources and changing network connectivity.
Innovation Solution
A phase identification algorithm using constrained k-means clustering of smart meter data, which extracts features from voltage time series and applies clustering techniques to accurately identify phase connectivity without the need for expensive equipment, leveraging network connectivity information to avoid mislabeling and reduce computational workload.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional phase identification methods are used, then equipment installation is required, but costs and device complexity increase
Solution Approach 1:
The patent replaces physical measurement equipment (phase identification devices, sensors, and hardware installations) with a computational algorithm that processes smart meter data. The phase identification is achieved through software-based constrained k-means clustering rather than mechanical or electrical measurement devices, eliminating the need for additional physical equipment while maintaining identification accuracy.
Solution Approach 2:
The patent uses existing smart meter voltage time series data as a proxy for direct phase measurements. Instead of installing dedicated phase measurement equipment, the system copies and analyzes the electrical characteristics already being captured by smart meters, extracting phase connectivity information from existing data streams without requiring additional hardware.
2Measurement precision
If comprehensive phase identification is performed across the entire network, then accuracy improves, but computational complexity increases
Solution Approach 1:
The patent segments the phase identification problem by first grouping smart meters into clusters based on voltage time series similarity, then identifying phases at the cluster level rather than individually for each smart meter. This hierarchical segmentation reduces the computational scope from O(n) individual identifications to O(n/k) cluster-level identifications, where k is the number of clusters.
Solution Approach 2:
The patent applies partial action by using constrained clustering that leverages existing network connectivity information to guide the phase identification process. Instead of performing exhaustive analysis on all possible phase configurations, the algorithm uses constraints derived from known network topology to focus computational effort only on plausible phase assignments, reducing the search space significantly.
3Reliability
If phase identification is performed frequently to capture changing network configurations, then reliability improves, but loss of time and computational resources increases
Solution Approach 1:
The patent performs preliminary action by pre-processing smart meter data to extract voltage time series features and organizing them for efficient clustering analysis. The constrained k-means algorithm pre-computes distance metrics and uses network connectivity information to establish initial constraints, so that when phase identification is triggered, the computational work is minimized and can be completed quickly even as network configurations change.
Data Source
AI summary
Techniques for phase identification using feature-based clustering approaches are disclosed. Embodiments employ linear and nonlinear dimensionality reduction techniques to extract feature vectors from raw time series. In an embodiment, a constrained clustering algorithm separates smart meters into phase connectivity groups. Another embodiment clusters smart meter data, where voltage measurements are collected from smart meters and a SCADA system. Then, customer voltage time series are normalized and linear or nonlinear dimensionality reduction is applied to the normalized time series to extract key features. Next, constraints in the clustering process are defined by inspecting network connectivity data. Then, a constrained clustering method is applied to partition customers into clusters. Lastly, each clusters phase is identified by solving a minimization problem. In another embodiment, a machine learning algorithm generalizes a subset of phase connectivity measurements to a distribution network, the algorithm being an extension of a Mapper algorithm in topological data analysis.