Phase identification in power distribution systems

Patent metadata

Field	Value
Publication number	US-2021285994-A1
Application number	US-201716323493-A
Country	US
Kind code	A1
Filing date	Aug 4, 2017
Priority date	Aug 5, 2016
Publication date	Sep 16, 2021
Grant date	—

Abstract

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.

First claim

What is claimed is:

1 . A method of phase identification, comprising:
receiving voltage measurements, the voltage measurements including a plurality of customer voltage time series; obtaining distribution connectivity information; normalizing the plurality of customer voltage time series by their respective standard deviations; defining constraints for a clustering process by inspecting the distribution connectivity information; applying constrained and unconstrained clustering to partition customers into a plurality of clusters; and identifying a phase of each of the plurality of clusters by solving a minimization problem.

2 . The method of claim 1 , wherein the receiving comprises receiving the voltage measurements from a plurality of smart meters in an electric distribution system.

3 . The method of claim 1 , wherein the receiving comprises receiving the voltage measurements from a supervisory control and data acquisition (SCADA) system.

4 . The method of claim 1 , wherein the obtaining comprises retrieving the distribution connectivity information from an electrical power utility operating an electric distribution system.

5 . The method of claim 1 , wherein the normalizing comprises applying one or more of a linear dimensionality reduction technique and a non-linear dimensionality reduction technique.

6 . The method of claim 5 , wherein the linear dimensionality reduction technique comprises applying principal component analysis (PCA) to extract key components from the plurality of customer voltage time series.

7 . The method of claim 6 , wherein extracting the key components comprises applying PCA on normalized customer voltage time series to extract the top n components as the key components.

8 . The method of claim 5 , wherein the non-linear dimensionality reduction technique comprises applying one or more of Sammon mapping, curvilinear components analysis, Isomap, and t-distributed stochastic neighbor embedding to extract key features from the plurality of customer voltage time series.

9 . The method of claim 1 , further comprising causing display of one or more of the identified phase of the plurality of clusters on a display device.

10 . A system comprising:
a computer-readable medium having instructions stored thereon, which, when executed by a processor, cause the system to:
receive voltage measurements, the voltage measurements including a plurality of customer voltage time series;
obtain distribution connectivity information;
normalize the plurality of customer voltage time series by their respective standard deviations;
define constraints for a clustering process by inspecting the distribution connectivity information;
apply constrained clustering to partition customers into a plurality of clusters; and
identify a phase of each of the plurality of clusters by solving a minimization problem.

11 . The system of claim 10 , wherein the receiving comprises receiving the voltage measurements from a plurality of smart meters associated with respective customers of an electrical power utility.

12 . The system of claim 10 , wherein the receiving comprises receiving the voltage measurements from a supervisory control and data acquisition (SCADA) system.

13 . The system of claim 10 , wherein the obtaining comprises retrieving the distribution connectivity information from an electrical power utility operating an electric distribution system.

14 . The system of claim 10 , wherein the normalizing comprises applying one or more of a linear dimensionality reduction technique and a non-linear dimensionality reduction technique.

15 . A non-transitory machine-readable storage medium comprising instructions, which when implemented by one or more machines, cause the one or more machines to perform operations for phase identification topological data analysis (TDA), the operations comprising:
obtaining data, the data including voltage measurements and training data; building, based on the training data, a forest of random trees; transforming each data point in the voltage measurements into a respective forest vector; running a t-Distributed Stochastic Neighbor Embedding (t-SNE) dimensionality reduction algorithm to produce numerical results for each data point; storing, on the non-transitory machine-readable storage medium, the numerical results in a table representing a filter function; building a base simplicial complex by inputting the training data into a Mapper algorithm; for each data point in the voltage measurements:
classifying new data by filtering the data point;
considering open sets in a covering that the data point lands in;
grouping the data point to a nearest trained cluster; and
labeling the nearest trained cluster.

16 . The non-transitory machine-readable storage medium of claim 15 , wherein building the forest of random trees comprises forming a hamming metric to learn a distance metric.

17 . The non-transitory machine-readable storage medium of claim 16 , wherein running the t-SNE dimensionality reduction algorithm comprises running the t-SNE algorithm with the hamming metric.

18 . The non-transitory machine-readable storage medium of claim 15 , wherein building the base simplicial complex comprises:
filtering the training data; inverting the open sets of the covering through the filter function; clustering inverse image subsets to form trained clusters; building a nerve of the covering; and causing display of a visualization of the base simplicial complex on a display device of one of the one or more machines.

19 . An apparatus comprising memory and processing circuitry coupled to the memory, the processing circuitry configured to:
obtain data, the data including voltage measurements and training data; build, based on the training data, a forest of random trees; transform each data point in the voltage measurements into a respective forest vector; run a t-Distributed Stochastic Neighbor Embedding (t-SNE) dimensionality reduction algorithm to produce numerical results for each data point; store, in the memory, the numerical results in a table representing a filter function; build a base simplicial complex by inputting the training data into a Mapper algorithm; for each data point in the voltage measurements:
classify new data by filtering the data point;
consider open sets in a covering that the data point lands in;
group the data point to a nearest trained cluster; and
label the nearest trained cluster.

20 . The apparatus of claim 19 , wherein building the base simplicial complex comprises:
filtering the training data; inverting open sets of a covering through the filter function; clustering inverse image subsets to form trained clusters; building a nerve of the covering; and causing display of a visualization of the base simplicial complex on a display device.

Assignees

• Univ California

Inventors

• Yu Nanpeng

Classifications

• H02J13/12

  Monitoring network conditions, e.g. electrical magnitudes or operational status · CPC title

• H02J13/10

  characterised by displaying of information or by user interaction, e.g. supervisory control and data acquisition [SCADA] systems · CPC title

• Y04S20/00

  Management or operation of end-user stationary applications or the last stages of power distribution; Controlling, monitoring or operating thereof · CPC title

• Y02B90/20

  Smart grids as enabling technology in buildings sector (smart grids supporting the management or operation of end-user stationary applications in general, or like technologies with no associated climate change mitigation effect Y04S20/00) · CPC title

• G01R29/18 Primary

  Indicating phase sequence; Indicating synchronism · CPC title