Condition monitoring platform for proactive and reactive operation and maintenance (O&M) with enhanced data analytic functionalities-Marios Theristis-University of Cyprus

Andreas Livera1, Marios Theristis1, George Makrides1, Juergen Sutterlueti2 and George E. Georghiou1 1PV Technology Laboratory, University of Cyprus, Nicosia, Cyprus 2Gantner Instruments GmbH, Schruns, Austria Condition monitoring platform for proactive and reactive operation and maintenance OM with enhanced data analytic functionalities Cyprus 212/12/2018 We are here Highlights of the PV Technology Laboratory 312/12/2018 Participation/coordination of over 60 national/international research projects Research funding of 18 MEuros over the past 10 years Comprises of 30 researchers Research quality/awards at international conferences Over 250 research publications Indoor/outdoor testing 412/12/2018 Testing site 512/12/2018 Official testing site for over 40 different manufacturers Introduction 612/12/2018 Key factor for future PV uptake is to reduce Levelized Cost of Electricity LCoE Increasing performance and reducing operating costs advanced monitoring Condition monitoring platform Failure detection and classification Data quality and sanity System health state Added Value Services Performance loss quantification Degradation rate estimation Background Objective 712/12/2018 Specific Objective Development of an innovative condition monitoring platform for proactive and reactive OM with enhanced data analytic functionalities Advanced baseline condition monitoring solution to ensure operational quality and optimise energy production Partners GI and UCY Project Innovative Performance Monitoring System for Improved Reliability and Optimized Levelized Cost of Electricity IPERMON [Solar-ERA.net project] Budget €400,000 Duration 36 Months Weblink http//www.pvtechnology.ucy.ac.cy/projects/ipermon/ State-of-the-art 812/12/2018 Visual inspection is the simplest method to detect visible failures The most popular technique for failure diagnosis is image analysis Methods based on advanced data analysis of electrical parameters are becoming increasingly popular Performance monitoring and data analytics 912/12/2018 Change from Descriptive analytics to Diagnostic/Predictive Analytics Adde dv alue Complexity Intelligent monitoring systems Intelligent data analytic features 1012/12/2018 Roust performance monitoring Intelligent data analytic features 1112/12/2018 Data quality and sanity Intelligent data analytic features 1212/12/2018 System health state Intelligent data analytic features 1312/12/2018 Failure detection and classification Intelligent data analytic features 1412/12/2018 Added Values Services Performance loss quantification Degradation rate estimation Platform functionalities – Data quality routines DQRs 1512/12/2018 Identify missing and erroneous data Estimate system availability and sensor deviations Correct data through data imputation techniques LOCF and linear interpolation Platform functionalities – PV system model prediction 1612/12/2018 Parametric and machine learning simulation models Highest prediction accuracy - FFNN Platform functionalities – Degradation Rate 1712/12/2018 Statistical and comparative techniques for trend extraction and estimation of the degradation rate Platform functionalities – System Health State 1812/12/2018 Comparative assessment between measured and predicted daily PV performance Classification of the relative error in ranked categories Comparative assessment between measured and predicted measurements against set threshold levels TL Statistical outlier detection rules 1912/12/2018 Platform functionalities – Failure detection Platform functionalities – Failure classification 2012/12/2018 Unsupervised procedures voltage/current/power ratio and fuzzy logic rules Supervised procedures k-NN, SVM, Decision and Regression Tress Failure patterns Bypass diode patternCurrent and voltage indicators Online Platform 2112/12/2018 Future 2212/12/2018 Digital Performance Architectures Digital Twin Concepts Int eroper abili ty IoTPhysical systems Virtual system Cloud Added services Grid control Performance analysis Forecasting Summary 2312/12/2018 PV performance measurements and analytical techniques are required to ensure optimal lifetime performance and to reduce LCoE Performance monitoring platforms consist of the Sensor network Data acquisition DAQ device Visualizations portal – Descriptive analysis Required accuracy and complexity depends on the PV system size and user objectives Future grid modernisation is the driver for advanced performance architectures 24 Acknowledgement Stimulating scientific excellence through twinning in the quest for sustainable energy TwinPV. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the agreement No. 692031 Together we do more for PV and Smart Grids Team 3 countries Over 100 Expert Researchers, Trainers One stop shop cells to modules to Grid Training, Testing, Research 26 Thank you for your attention 27 More information Websites www.pvtechnology.ucy.ac.cy https//www.gi-cloud.io Dr. Marios Theristis PV Technology Laboratory University of Cyprus Email theristis.mariosucy.ac.cy