To achieve the EU’s climate change and energy policy objectives by 2030, it is not only necessary to build new power lines and stations, but it is essential to exploit the power systems by integrating information and communication technologies (ICT), to build controllable microgrids and accelerate the realization of the smart grid.
The SMART MICROGRID DEVELOPMENT TECHNOLOGY BASED ON DIGITAL TWIN (SMARTECH) project is an R&D&I project financed with the support of grants awarded by Iceland, Liechtenstein and Norway through the EEA financial mechanism 2014-2021, within the “Romanian Energy Program”. Until now, the consortium enjoys a visibility in the scientific framework, the project having the following intermediate results:
- Sizing model of microgrid components and performance calculation (number of photovoltaic panels, used area, number of batteries, park power, inverter power, ROI, etc.), using Matlab
- Development of the automatic reporting tool in Word format of the technical specifications resulting from the component dimensioning model and the integration of microgrid performances
- Development of algorithms for balancing production and consumption under energy market conditions (Time Shifting, Peak Shaving, Cuckoo Search), using Matlab
- Development of the tool for contextual analysis of predictive maintenance indicators, based on Machine Learning algorithms, which includes both basic statistical data, such as mean or median, but also complex functions developed based on data taken from the real case study. The dispersion was analyzed, as a value corresponding to a statistical series, which measures the distance of each value of the series from the sample mean, but also the standard deviation, used in descriptive statistics especially to define intervals in which the vast majority of observations are found. In the example provided for support, indicators such as confidence interval, Skwness (represents the measure of the deviation of the given distribution from symmetry) were also calculated. If the skewness is non-zero, then the distribution is skewed and will be “tailed” either to the left or to the right. The “sharpness” of a distribution was also analyzed, in the sense in which the segment or population is disposed to have extreme values (outliners), the scatter diagram, the correlation coefficient, the level of significance, etc. An analysis intended for Machine Learning techniques such as survival analysis and statistical analysis of residuals were developed.
- 5-screen platform for viewing fault/health parameters:
- Parameter view screen (temperature, active power, State of Charge, remaining power to discharge, Power Setpoint, etc.) for different components of the microgrid
- The screen with graphs for the number of alarms in 24h, the consumption forecast for 24h, the production forecast for 24h, etc.
- The equipment specifications screen is intended for microgrid components, having the headings: specifications, attachments, maintenance history and observations. The attachments represent the entirety of the operation manuals, CAD schematics, historical data, installation manuals and CE certificates associated with each component’s equipment.
- The overhauls centralization screen contains two tables that centralize unplanned overhauls and the overhaul scheduling table.
- Predictive maintenance KPI dashboard screen showing the RUL value, in days, represented both numerically and on semi-circular indicators with modifiable lower and upper limit values. The percentage values for availability, performance, quality and OEE will be displayed on gauges in the form of a dial.
For more details, you can contact the communication manager: email@example.com