Due to the role shifting in the smart meter system, the energy industry its own is at a digital transition period. Machine learning-enabled smart meter data analysis is still at its early stage. There are several key questions that haven’t been answered yet. For example, given a historical data on consumption from end user, how to forecast the future consumption trends while taking users’ difference into consideration, what are the inner and outer drivers that hiddenly decides on consumption patterns or signatures from end-users, and what events could be trigged by such consumption behaviors and their effect on the smart grid system, has not been exploited yet. The solutions to those questions are, actually, the key technologies in the smart grid system in this new AI era, and we have listed two major technologies that we aim to solve during this project, that is: 1) propose new machine learning models to help forecasting the energy usage and corresponding suggest new explanation mechanism to interpret the model; 2) Analysis of end-user on external factors and their impact on energy consumption, aiming to disclose how user behavior affects economically with limited data.
Our project in the long term aim to seek external funds to support our research on shaping consumer behavior and DSO’s investments. Consumer activities account for a large portion of the total energy demand. Therefore, a deep understanding concerning various factors on the consumer side would bring new insights and possibilities for utility companies, governmental agencies, and various entities with environmental concerns aiming at lowering and shaping energy consumption patterns, and finally to achieve peak load reduction, load smoothing, and hence carbon emission curtailment. In turn, the project outputs in the long term have the potential to address the goal 7 in UN Sustainable Development Goals (SDG) to expand the use of renewable energy beyond the electricity sector with sustainable, reliable energy, by supporting the demand from the consumer side and urging the needs of renewables to the producer side.
