Fundamentals of CIM for big data integration and interoperability Grid Analytics Europe 2016 5-6 April 2016 Ivo Kuijlaars 06 April 2016 Stefan Pantea, Interoperability and Systems Engineering Expert, National Grid Nejc Petrovič, Technical Researcher, Elektro Gorenjska Ivo Kuijlaars, Consultant, DNV GL 1 SAFER, SMARTER, GREENER
Introduction Big data integration and CIM Ivo Kuijlaars 25 minutes CIM fundamentals, TSO experience Stefan Pantea 25 minutes DSO experience Nejc Petrovič 25 minutes Stefan Pantea All Demo Discussion 30 minutes 15 minutes Big data integration and CIM: Big data requirements Big data situation Big data integration and interoperability Phase 1: batch based Data management architecture styles Mapping of data Big data integration and interoperability Phase 2: service based Common Information Model Interface Reference Model CIM as a semantic model CIM as a reference model CIM for unique identification CIM pros and cons 2
About DNV GL In the energy industry, DNV GL delivers world-renowned testing and advisory services to the full energy value chain including renewables and energy efficiency. Our expertise spans onshore and offshore wind power, solar, transmission and distribution, smart grids, and sustainable energy use, as well as energy markets and regulations. Helped realizing reliable grids for 90 years Over 2,500 employees in energy Experts with in-depth knowledge Integrated services Risk-based verification concept 3
Big data requirements For analysis the following data requirements are often mentioned at utilities: 1. Single definition of data 2. Single Point of Truth: no redundancy 3. Integration of distributed data 4. Data lifecycle 5. Data lifetime: system independent 6. Data quality: complete, right and up-to-date 7. Data retrieval performance Plan Remove Maintain Operate Data lifecycle Design Build What is the current situation? 4
Big data situation Recognisable situations at utilities: Data is stored in silos Only system IDs: no 'Globally Unique Identifier' Data governance is poor Data definition is not harmonised Data reference model is not shared Poor data quality control Intensive search needed to acquire needed information Intensive maintenance work on data quality and data interfaces Poor data: poor decisions! Governance? Do you really want this? How to merge data for analysis? 5
Big data integration and interoperability Phase 1: batch based Data analysis on integrated data, ideally from a central register Asset data is stored in different systems: EMS, DMS, OMS, GIS, ERP, CAD, ECM, etc. These have different data models and different data definitions. However data can be merged through a staging process. Mapping is crucial here. Are there alternatives for a central register? 6
Data management architecture styles How to realise data mapping? 7
Mapping of data Interfaces can be realised in 4 ways: 1. Custom interface: every interface needs to be developed and maintained 2. Mapped interface: a customised reusable interface based on a reference model 3. Multi purpose adapters, based on a reference model 4. Matching physical data models: no need for an interface What is an adapter? 8
Big data integration and interoperability Phase 2: service based Once data is related or merged into a central register, the next phase are the real-time interfaces between source systems and the central register. These use adapters to map the data to the central register, but also directly to the other systems Mapping needs a reference model. Which one? 9
Common Information Model The CIM is an international IEC standard, now globally accepted for modelling the information exchanges required in electric utilities. The CIM is independent of any individual application, middleware, or message protocols used for data exchange. The interoperability enabled by the CIM standards is a key factor for achieving the Smart Grid vision. The CIM is defined in WG13, 14 and 16. Smart Grid Architecture Model (SGAM) How is the CIM structured? 10
CIM Interface Reference Model The IRM provides a framework to identify information exchange requirements (use cases) for business functions. CIM now has over 600 classes with thousands of attributes and associations. IRM helps to prevent duplication of data definitions in CIM. Interface Reference Model (IRM) What are data definitions in CIM? 11
CIM as a semantic model Back to requirement 1: Single definition of data CIM is a semantic model: it describes the components of a power system. This provides consensus on the interpretation of each class and attribute, removing ambiguity and duplication of definitions. Examples: Recloser: Pole-mounted fault interrupter with built-in phase and ground relays, current transformer (CT), and supplemental controls. Sectionaliser: Automatic switch that will lock open to isolate a faulted section. It may, or may not, have load breaking capability. Its primary purpose is to provide fault sectionalising at locations where the fault current is either too high, or too low, for proper coordination of fuses. How can you use CIM for mapping? 12
CIM as a reference model cla ss DCIMPhaseModel Back to mapping. CIM is a common model to which all physical data models can map. This enables interoperability. Cor e:: ConductingEquipment Wir es:: Ener gyconsumer +EnergyConsumer Wir es::regulatingcondeq + controlenabled: Boolean [0..1] 1 W ir es:: Shunt Compensa tor +ShuntCompensator 1 Cor e:: Equipment 0..* +ShuntCompensatorPhase 0..* +EnergyConsumerPhase W ir es:: Energy ConsumerP hase + pfixed: ActivePower [0..1] + pfixedpct: PerCent [0..1] + phase: SinglePhaseKind [0..1] + qfixed: ReactivePower [0..1] + qfixedpct: PerCent [0..1] W ir es:: ShuntCompensa tor P ha se + maximumsections: Integer [0..1] + normalsections: Integer [0..1] + phase: SinglePhaseKind [0..1] IdentifiedObject Cor e:: Power Sy stemresource Wir es:: ACLineSegment +ACLineSegment Wir es::aclinesegmentphase Wir es:: Conductor 1 0..* +ACLineSegmentPhases + phase: SinglePhaseKind [0..1] +ACLineSegment W ir es::cut 1 +Cut 0..* + lengthfromterminal1: Length [0..1] W ir es::switch 1 +Switch 0..* +SwitchPhase Wir es::switchpha se + closed: Boolean [0..1] + normalopen: Boolean [0..1] + phaseside1: SinglePhaseKind [0..1] + phaseside2: SinglePhaseKind [0..1] W ir es:: Disconnect or W ir es:: Sectiona liser Wir es::fuse What about a world wide ID for all components? Wir es::jumper Wir es:: Gr ounddisconnector W ir es:: P rotectedswitch Wir es:: Recloser Wir es:: Loa db r ea k Swit ch Wir es:: B r ea k er 13
CIM for unique identification Three CIM pillars to start with data integration: Single (semantic) definition of data Data mapping to a common model Unique identification CIM (part 454, Business Object Registry Service Specification) models shared identification of objects, with an mrid of GUID. Systems can keep using their system IDs and local names. With this the CIM model offers a working registry service to relate data in silo s. What are the benefits of CIM? 14
CIM Pros and cons Advantages: Proven, mature and direct usable standard: CIM provides common semantics for information exchange between heterogeneous systems Fewer interfaces: reduces data engineering efforts and maintenance on current interfaces Consistent interfaces: easier understandable and maintainable/expandable Improved data quality: Unity in definitions, Less errors by clear definitions and validation of the data No vendor lock-in Uses industry standard modelling notation (UML, XML) Good architectural principal: Strive for unity and standardisation Disadvantages: First interface is a lot of work CIM may appear more difficult, when the approach is for a point-to-point interface solution Conclusion: CIM enables big data integration and interoperability 15
Thank you for your attention Ivo Kuijlaars Ivo.Kuijlaars@dnvgl.com +31 26 3562668 www.dnvgl.com SAFER, SMARTER, GREENER 16