MALENA: Machine Learning System for Energy Data Analysis and Management
The aim of MALENA is to develop innovative and state-of-the-art software tools covering the participation needs of the Public Power Corporation (PPC) of Greece in the Day-Ahead and Intra-Day energy markets of the European Target Model, introduce innovation and expertise within PPC, and provide PPC with a software solution that sets the company free from license restrictions and third-party software companies. At the same time, the provision of a personalized service to consumers, giving them integrated access to their energy data and allowing them to manage their consumption, broadens PPC’s services portfolio with the integration of current trends in the energy sector. In order to pursue all these objectives, MALENA investigates two main research paths for forecasting multiple future values of time series: a) deep neural networks, b) multi-target prediction.
On one hand, deep neural networks have revolutionized machine learning during the past years achieving top results for unstructured data (images, video, audio, text). On the other hand, multi-target prediction techniques allow the exploitation of algorithms – such as the extreme gradient boosting algorithm, which achieve top results in tasks with structured data involving multiple target variables, such as load forecasting. The project studies the application of the aforementioned techniques in time series of energy and weather data. In addition, the project studies graph mining techniques and methods for scaling up machine learning to data streams using graphics processing units and cloud computing.
Results: After a detailed study of end user requirements and the respective specifications, the project generated an integrated software prototype with a user-friendly web interface that can collect data (power generation and consumption data, meteorological data) and provide a) continuous load forecasts b) personalized load management and c) renewable energy sources power generation forecasts. The system as a whole as well as its sub-systems were thoroughly evaluated with respect to their functionality and results. Moreover, novel deep learning methods were introduced and some of them were incorporated in the prototype software. More specifically, novel methods were developed for short term and day ahead electric load demand forecasting as well as wind energy generation day ahead forecasting. The methods utilized diverse approaches such as tree-based ensembles, lightweight neural networks, anchored input-output learning, online self distilation, residual error learning etc. The research outcomes of the project were disseminated through in 11 scientific conference papers.
The project has been Co‐financed by the European Regional Development Fund of the European Union and Greek national funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the call RESEARCH – CREATE – INNOVATE (project code: Τ2EDK-03048)
