Load Forecasting Using Neural Networks and Blockchains for Low Voltage Distribution Networks

Abstract

This work presents an efficient and reliable short term load forecasting algorithm that can determine the optimal operating conditions for end consumers. The proposed solution combines smart meters readings from the electrical distribution company along with weather conditions, and previous consumption figures and feeds them into a reliable neural network-based load forecasting and scheduling algorithm to ensure optimal operating conditions. Predicted power consumption figures will be presented to end users to plan for optimal energy usage. Furthermore, A Blockchain-based communication system is used for information exchange throughout the network. End users will have their actual power consumption figures sent to the utility provider to ensure the planned operation and efficient use of power plants and implementation of dynamic pricing.