Energy in Africa

The Case:

Standard Microgrid is an innovative tech company that is aiming to reinvent power. Standard Microgrid has been providing renewable power to communities in Sub Saharan Africa for the past 4 years.

With a projects in Zambia, Tanzania, Kenya and the DR Congo, Standard Microgrid is poised to become the market leader in smart digital distributed renewable energy services in Africa.

Over the past 4 years Standard Microgrid have collected vast amounts of log data from their smart meters. These logs contain valuable information regarding the operation and status of each device. Netlyt's team was tasked to help identify the causes for switch malfunction. And to help identify any faulty devices which might be affecting customers on Standard Microgird's network.

The Challenge:

Standard Microgrid's switches recorded a log entry every 15 minutes for the past 4 years. This produced a large log file with hundreds of thousands of entries. Far too large to be analyzed by traditional means such as using Excel spreadsheets 5or manual inspection. Some of the malfunctions produced corrupt entries which further complicated the task. Techniques developed for machine learning and Big data were required to deal with the vast amounts of information and wide range of errors.

The Solution:

The team at Netlyt used a wide variety of techniques and tools that we've developed in our work to extract as much information from the corrupt entries as possible. We've managed to identify 9 different types of errors, clean and visualize the retrieved information. We've identified some of the switches that were exhibiting odd behavior (turn off when they should be on). Using our tools we've managed to correlate the data from several log files in order to identify the most common events that caused log corruption.