Model-driven software design for smart grid data analytics

Practical data analytics for the smart grid requires a software platform which enables the distribution companies to better correlate the projects, improve understating of system requirements and simplify system design by decomposing its complexity and large-scale data. This paper proposes the model-driven software design (MDSD) to managing and optimizing the smart grid data. It defines a language for visualizing, specifying, analyzing, and documenting the distributed object-oriented data. An MDSD experience is reported that employs three software environments interfaced with each other in order to create a chain of the desired methods for an energy efficiency program. The aim is to transform the data into actionable decisions accessible and understandable by distribution companies by taking software development into a higher level of abstraction. As a pilot project, a comprehensive data logging were performed for an MV/LV distribution system. Subsequently, the technical losses were studied using the developed software package. A Java-based GUI, MATLAB/Simulink, and an embedded C-based data management and calculation module are used in the proposed MDSD.