Your personal signature is a means to legally identify yourself, whether it is on a contract agreeing to perform a specific action or on a check promising to pay the other party a specified amount. The use of a signature was first noted in ancient Egypt, followed by ancient Greece and Rome.

Modern use began in 1677 with the State of Frauds Act passed in the English Parliament, which stated that contracts must be signed as a measure of guarantee against fraud. One of the most notable signatures is probably John Hancock’s on the Declaration of Independence. The written signature (and even cursive writing) has given way to electronic or digital signatures since the 2000 E-Sign Act legally validated electronic contracts, paving the way for eSignature technologies in businesses around the world.

Signatures and PQ

You might be asking, what does this have to do with power quality? PQ signatures have been a topic in the community since the first PQ monitors had graphic capability, capturing and displaying voltage and current waveforms. The “Handbook of Power Signatures” and “Power Quality Analysis” books, published in the late 1980s, had hundreds of pictures of waveforms and explanations written by the experts as to what caused them, and how to minimize their effects. 

Other reference texts followed, but were mostly written by experts for other experts in the field. NJATC undertook a revised approach with its joint venture with Dranetz in 2010 to produce a PQ Analysis book for electricians, which is still published today.

The process with PQ signatures was to review those references and compare the voltage and current waveforms from those captured by a PQ monitor at the job site to see if they matched to help determine what caused the disturbance. Research projects, especially at universities with PQ laboratories, have been working on algorithms to automate this project, turning it from signatures analyzed by humans to ones analyzed by electronic systems. 

Advanced techniques such as wavelets and neural networks are just a few of the many techniques applied in this quest. In 1994, a software program called AIPower was designed to do such, providing a “probability of correctness” factor to its answers. But humans have this tendency to think that getting it proven wrong once is enough for some to see no value in it.

Developing standards

Five decades after the first graphical PQ monitors, it seems that the PQ signature is still analyzed mostly in the domain of experts. Having PQ monitors or software that can state—with a guarantee—that the 0.81 PU voltage sag was caused by a single line to ground fault on phase A conductor due to a red squirrel with paws touching the conductor and a tail touching a grounded conductor is still elusive. Progress is still being made in that direction. 

The IEEE PES PQ Subcommittee formed a Working Group on Power Quality Data Analytics a few years back. Its first action was to produce the task report “Electric Signatures of Power Equipment Failures (TR73)” in 2019. The 86-page report has numerous examples for common PQ disturbances, with their signatures and explanation of what failed and why. 

IEEE P3139 officially began its four-year journey in 2021 to produce a guide, with the following scope: “This guide provides an overview of new and existing methods to extract meaningful information from power quality (PQ) data using data analytics. PQ data may include value logs (time series), waveforms (point on wave), phasors, spectrums, and characteristics recorded during PQ disturbances such as voltage sags, voltage swells, transients, rapid voltage change events, and metadata related to PQ data. 

“Analytic techniques such as feature engineering, transformation, and machine learning are presented. Pre-processing of data includes methods for data collection, wrangling, structuring, and data quality. 

“Effective methods of presentation and reporting of data are covered, including charting, geospatial, and other visualizations. Case studies are provided demonstrating examples of PQ data analytics. Roles of personnel managing PQ data analytics are included. These can include PQ subject matter experts, data engineers, data analysts, data scientists, and end-users/stakeholders.”

The group is actively looking for data contributions to improve the document, and this is a case where more will make it better. Those wishing to contribute should access the following URL: grouper.ieee.org/groups/td/pq/data/downloads/How_to_Access_and_Contribute_to_Data_Collection.pdf