Grid Diagnostic deployment Sample Clauses

Grid Diagnostic deployment. The main objective of this task is to devise predictive models, algorithms and tools that help to guarantee grid stability and resilience through a deep knowledge of the effects in the grid of a highly variable generation and demand (e.g. renewable and electric vehicles), faults in the grid facilities, faults in the distribution grid or attacks. A Reliability Centred Maintenance (RCM) approach was utilised to evaluate and assess failure modes of assets and their components, which could be detected before functional failure occurs and the consequences of these failure modes. A multitude of applicable failure modes were identified, however the sensor inputs to the diagnostic module were limited to only sensor technology currently incorporated on the MUSE GRIDS system. The following failure modes were identified as suitable for development in the Osimo site: • Battery Analysis of the ▇▇▇▇ • IGBT Analysis of the ▇▇▇▇ • Voltage related failure of the PV modules As not previous data of fault modes is available, it was decided that the best approach for the development of the algorithms is through data driven rule-based approach. Anomaly detection algorithms are being developed based on the real time data which will identify outliers that are an indication of a potential fault in the system. Such algorithms have the advantage of being less complex with limited computational power, thus can be easily deployed within the controllers compared to more complex machine learning techniques. Also, due to the lack of historical data the idea of the development of machine learning techniques was abandoned. Machine learning techniques pre-require substantial amount of historical data in order to learn patterns that are representative of system faults. Machine learning techniques approach could be used once enough representative data that describes the behaviour of faulty system is gathered.