Powering ‘Energy as a Service’ for a leading Off Grid Solar provider

  • Tuesday, January 01, 2019
About Client
  • A Fortune 500 company offering Off Grid Solar solutions
  • Spread over African and Indian markets
  • World leaders and innovators in solar lighting solutions for both consumers and cities

Business Problem

The off-grid solar solutions were targeted to replace expensive diesel generators and provide a greener, cheaper and reliable option. They were also exploring a new commercial model move from capital expenditure to charging for energy consumed.

Three key aspects needed to be addressed for the new commercial model to be successful:

1. Since, remote healthcare units were one of the target customers, accurate forecasting of energy production was critical. Because loss of power means loss of lives

2. Over-engineering the panel size was to making it economically nonviable

3. Poor battery performance was a showstopper to provide reliable product, thus impacting customer satisfaction

How Cerebra solved the problem?

Cerebra ingested the following key parameters every minute from the off-grid solar solution

ï‚· Solar Panel Parameters – Solar Panel Voltage, Solar Panel Current, Solar Panel Power, Solar Panel Energy

ï‚· Battery Parameters – Battery Voltage, Battery Current, Battery Power, Battery Energy +ve & -ve

ï‚· Streaming sensor data such as irradiance sensor and panel temperature sensor

ï‚· Weather data such as temperature, humidity, and rainfall

solaroem

Using the raw data, several rich vectors were generated such as Solar Irradiance Efficiency, Discharge Velocity, Survival probability. Arimax Forecasting Algorithms were implemented to forecast accurate energy production. A combination of regression models was implemented to monitor and diagnose asset related health.

Scale

  • 80+ data points every minute streaming several off-grid deployed solutions globally
  • Over a period of 1 year, it resulted in several billion data points
  • Onboarding of a new solar solution was a matter of minutes, thus helping rapid scale up.

Impact on outcome

  • They identified several locations where the installed capacity was over engineered by 76%
  • 93% accurate power generation forecast lead to more reliable solution
  • Proactive attention to the abnormal behavior of battery parameters, helping identify poorly performing batteries and helping repair / replace before a significant drop in performance.


SUBSCRIBE TO Flutura

Join our mailing list to stay up to date on the latest developments in the IIoT space.