DECEMBER 16, 2020
Machine learning takes dLab’s grid insight platform to next level
Energy technology company Dlaboratory Sweden AB (dLab), based in Lund’s innovation center Ideon, has in the past few years grown considerably in sales and market outreach. That has however, not stopped the company’s R&D team to continue developing new applications and solutions for the customers. A development that is now leading dLab into the realm of new technology, namely machine learning.
- There is so much improvement to be made through using machine learning; we are very well positioned to embark on this endeavor – there is really no end to what we could do in grid analysis if we could find new patterns in all data. And machine learning would do that for us, says Kewin Erichsen, development manager at dLab.
dLab’s grid analytics platform is based on dAnalyzer, a unique patented algorithm, automatically analyzing data in substations. Harnessing this data, operations can make faster fault detection, get early warnings for imminent faults, and work proactively with maintenance.
Adding machine learning to the compiled data would mean that dLab’s solution could, for instance, provide users with a more detailed categorization of faults. The warnings for earth faults could then also indicate the type of incident causing the disturbances, e.g., rough wind, falling trees or other, and what kind of component was affected and where.
dLab’s development team does not only explore the advantages of machine learning but works on features to improve and broaden the use of dLab’s grid insight platform. One aspect of broadening the solution is to use it to implement condition-based maintenance in substations. A recently completed project with Kraftringen for Energiforsk is the basis for that line of thought. Another project opens possibilities in exact localization through combining dLab’s solution with network information solutions. Now, dLab is looking at ways of turning the findings into products for the customers.
- All our development is actually about the customers. While it might be pure fun for us in the development team, it is really all about making our platform more useful for our customers. Which is a very engaging mission, and we are proud to step up to deliver on that, says Kewin Erichsen.
The next new thing for dLab might however be something quite different; establishing a health index for grids. This could lead to interesting insights for grid owners by for example cross referencing the index with costs to form a basis for maintenance decisions. With the development of the health index, dLab will not only be able to provide energy distribution companies with early warnings for imminent faults, but also give a reference point for long term grid investments.