Automated analysis of power quality measurements by using powerful pattern recognition methods

Important results from the project

The purpose of this project was to transfer new knowledge within artificial intelligence (AI) to Metrum. The aim was that Metrum during 2017 introduces their first product based on pattern recognition for automatic detection and classification of power network disturbances. Manual analysis of measurement data is thereby replaced by pattern recognition algorithms, developed in this project. The goal fulfilments of the project are good since Metrum intends to introduce their first product based on the results from this project during 2017.

Expected long term effects

Metrum has gained significant new knowledge within AI which has resulted in new software, based on pattern recognition, that replace manual analysis of power quality disturbances with an automatic analysis. Expected effects of this are more reliable analysis in combination with released time for the user that can be used for other tasks. All this together results in increased customer value and significantly strengthened competitiveness of Metrum. In addition, long-term cooperation has been established between the University of Borås and Metrum, which benefits both partners.

Approach and implementation

The project began with activities that provided Metrum´s employees knowledge within AI that was used to replace manual analysis of measurement data. Thereafter, a general development platform was developed based on Python. This platform was then used in the development process of the final software, which shows high accuracy. The project was concluded with the writing of a scientific article describing the project results. The article was presented at Metrum’s international sales conference in May 2017.