Item -Wise Billing For Household Appliances Using Deep ML/AI

About Client

An energy-tech company based in the US that partners with B2B distributors to generate electricity bills for households

Challenge

There were 3 broad challenges faced by the client:

  • Utility distributors were unable to understand the energy leakages and opportunities for energy optimization in their region
  • End customers were having disparity with the monthly utility bills
  • Malfunction of smart meters were detected manually by the client which resulted in loss of time and resources of the distributors

Our Solution

  • Deep Machine Learning Algorithm which had Internet Of Things (IOT) sensors to check for any malfunction or errors in the smart meters was applied after training on 10M+ data points
  • The algorithm learnt what appliance was used at what time of the day with visual representation of graphs, for e.g- Television usage was identified with a unique wave while Air Conditioning consumption had a different wave pattern
  • This helped in accurately generating the end of the month utility bills for households
  • Additional support was given in generating email notifications to the end consumers about their appliance-wise usage at regular intervals
  • Email notifications alerted the consumers to act more consciously thereby, strengthening consumer relationship
  • The Deep Machine Learning system subsequently helped the distributors to identify the potential customers based on third-party data and optimise the energy at their regions

Outcome

  • Itemized bills were generated for each household’s electricity consumption to eliminate any kind of disparity
  • Email Notifications are sent to utility consumers for informing their consumption at regular intervals which helped them to track and use energy consciously
  • Utility distributors could optimize energy of their regions effectively by analyzing the energy leak
  • AI-powered alerts were generated in case of failure or malfunction of smart meters to the client

Leave a Reply

Your email address will not be published. Required fields are marked *