Web based supervisory control and data acquisition Thomas

Web based supervisory control and data acquisition Thomas Dreyer (RACOS) David Leal (CAESAR Systems) Andrea Schröder (FGH) Michael Schwan (FGH) 2 nd International Semantic Web Conference, October 2003 http: //www. scadaonweb. com

Scada. On. Web Project partners FGH - Forschungsgemeinschaft für Elektrische Anlagen und Stromwirtschaft e. V. Bacher - Bacher Consulting Caesar - Caesar Systems Limited Cygnus - Cygnus Engineering AG LABEIN - Fundacion Labein RACOS - Racos Technische Informationssysteme SEf. AS - Sintef Energy Research 3 3

Our objective To put SCADA data on the web ð So that SCADA data can be handled by cheap software ð Low cost applications for which dedicated SCADA systems are not affordable The opportunity ð A semantically precise SCADA object ð Open source software to access it defined by the project written by the project ð Validated by project demonstrators 4 4

Requirements for the technology SCADA systems are different because they handle large amounts of data - gigabytes The data is structured into arrays of different dimensions ð for each time point there is a 1 D array of 1000 measurements ð 10000 time points gives a 2 D array (10000 1000) Efficiency requires ð maintenance of the array structures - no ‘confetti’ data ð numbers held as binary format Use over the web requires ð ð 5 XML for the semantics of the data ontologies for the measurement process - e. g. sampling rates ontologies for physical properties, physical quantities and units ontologies for measurement quality 5

Structured semantic data block gigabytes of structured data extract and download subsets variables positions states/times semantics acknowledgements to 6 6

What we defined describing OWLself semantic wrapper measurement object • compact binary data block • full semantics HDF 5 binary data • measurement quality system definition property ontology measurement quality ontology units ontology 7 OWL 7

HDF 5 (Hierarchical Data Format - version 5) Standard for the exchange of structured binary datasets ð developed by NCSA ð used for NASA and ESA for satellite telemetry ð used for oil exploration data Exchange of binary data between computers ð access library availably on many different types ð callable from C, Fortran, and Java Range of tools for ð data addition and extraction; ð display and editing ð interface to and from XML But the HDF 5 tools do not know about the meaning of the data 8 8

Demonstrators · Flood warning system based upon many remote sensors · Flexible metering of domestic and small industrial consumers · Data sharing within a balance group in the energy market · Control of distributed wind and hydro electricity generation · Condition based maintenance of remote equipment 9 9

Brig, Switzerland, September 23, 1993 © C. Heinen, 1993 10 10

Brig, Switzerland, September 1993 © G. Escher 1993 11 11

Proposed Mac. Eco Web interface 12 12

Different presentations - graphs 3 different weeks 3 different mondays 13 13

Wind park main display 14 14

Wind mill 15 15

Monitoring data of transformers Failure rates of transformers with tap changers: (CIGRE failure statistic) 16 16

Fundamentals of the technology Formal approach to engineering data ð based upon ISO 15926 ð a design is a Class Physical property, quantity and scale ð a physical quantity is a member of a physical quantity space ð a physical quantity space is the domain of a scale Product, activity and state ð 4 D world view Measurement quality ð quality is a measurement of the measurement system Distribution ð a distribution is a Property Descriptions of a distribution ð described by a numerical table ð stored as HDF 5 17 17

Engineering use of classes http: //www. scadaonweb. com

Designs are classes cast iron engine part <Class ID=“Rover 75 Crankshaft”> <sub. Class. Of resource=“. . . #V 6 Crankshaft”/> <sub. Class. Of resource=“. . . #SGCast. Iron. Crankshaft”/> part <sub. Class. Of rotating resource=“. . . #Part. With. Oil. Holes”/> crankshaft </Class> SG cast iron crankshaft V 6 crankshaft unmachined-surface part Rover 75 crankshaft SG cast iron grade XYZ part with oil holes Rover 75 crankshaft rev 1. 2 19 19

‘Members’ of a design Rover 75 crankshaft rev 1. 2 member of class The crankshaft in test T_1234 <Product ID=“crankshaft. In. Test. T_1234”> <type resource=“. . . #Rover 75 Crankshaft. Rev 1. 2”/> </Product> 20 20

Definition of design classes length in range 9. 99 to 10. 01 m XYZ_123 material S_1234 <Restriction ID=“ 10 Metre. Widget”> <on. Property resource=“widget. Length”/> <Class ID=“XYZ_123”> <all. Values. From resource=“ 10 Metre. With. Tolerance”/> <intersection. Of parse. Type=“Collection”> </intersection. Of> <Class about=“ 10 Metre. Widget”/> </Restriction> <Class about=“S_1234”/> </intersection. Of> </Class> 21 21

Physical property, quantity and scale http: //www. scadaonweb. com

Physical quantity as a class The class called “ 10 Kg” has instants in the lives of material objects as members. An instant in the life of a material object is a member if and only if it has 10 times the same inertial mass as the reference lump of platinum in Paris. <Product. At. Instant ID=“My. Widgetat 20031021 T 10: 30”> <type resource=“Mass. Of 10 kg”/> </Product. At. Instant> But usually it is more complicated: Vessel Physical object 23 mass. When. Empty Mass <Product ID=“my. Vessel”> Physical property Physical <mass. When. Empty resource=“Mass. Of 10 kg”/> </Product> quantity 23

Getting units right A unit or scale is a Property of a physical quantity. Mass kg Real <Class ID=“Mass. Of 10 kg”> <kg> <Real> <decimal>10. 0</decimal> </Real> </kg> </Class> 24 24

Getting units right A unit or scale is a Property of a physical quantity. Vessel mass. When. Empty Mass kg Real <Class ID=“my. Vessel”> <mass. When. Empty> <Mass> <kg> <Real> <decimal>10. 0</decimal> </Real> </kg> </Mass> </mass. When. Empty </Class> 25 25

Measurement quality http: //www. scadaonweb. com

A measurement system at an instant physical system at instant physical system instant. Of part. Of measurement system at instant measurement system instant. Of measured property measured system quality 27 27

IEC 61850 Measurement quality Space Validity Allowed values Good Questionable Invalid Detail. Qual 28 Overflow. Status Overflow Not. Overflow Out. Of. Range. Status Out. Of. Range Not. Out. Of. Range Bad. Reference. Status Bad. Reference Not. Bad. Reference Oscillatory. Status Oscillatory Not. Oscillatory Failure. Status Failure Not. Failure Old. Data. Status Old. Data Not. Old. Data Inconsistency. Status Inconsistent Not. Inconsistent Default Substituted Source Process Test. Status Test Not. Test Operator. Blocked. Status Operator. Blocked Not. Operator. Blocked 28

Temperature, time and validity <Product. At. Instant> <instant. Of. Product resource="#my. Measurement. System"/> <measured. Temperature> <Thermodynamic. Temperature> <celsius> <Real><decimal>27. 5</decimal></Real> </celsius> </Thermodynamic. Temperature> </measured. Temperature> <measured. Time> <utc>. . </Time> </measured. Time> </utc> <measured. Validity> <Validity about=”. . . /iec 61850 Quality. owl#Good"/> </measured. Validity> 29 </Product. At. Instant> the temperature the time the validity 29

4 D modelling http: //www. scadaonweb. com

Things in the real world space Product its identifier what type it is how it is connected 31 Product Life Segment (activity) when: from --> to maximum flow rate total flow average flow rate Product At Instant (state) time when: at flow rate at instant 31

Distributions sampling decomposition flow rate total flow in period time 32

Property distributions with respect to space and time http: //www. scadaonweb. com

Array of sensors at an instant <Property ID=“measured. Temperature”> <domain resource=“. . . #Sensor”/> <range resource=“. . . #Thermodynamic. Temperature”/> temperature </Prperty> <Property ID=“my. Distribution”> <sub. Property. Of resource=“#measured. Temperature”/> <domain resource=“. . . #My. Sensor. At. Instant. Set”/> </Property> For each sensor at an instant in the set there is a temperature my sensor at instant set subclass sensor at instant 34 my distribution subproperty temperature measured. Temperature 34

Array of sensors at an instant temperature For each sensor at instant in the set there is a temperature and quality my sensor at instant set my distribution temperature and quality space parameterisation: numbering of the sensors scale: Celsius and representation of quality by an integer vector For each integer there is a real and quality identifier 1 to 20 35 array real and integer vector 35

Semantics and mathematical description product (instrument) sampling or decomposition domain of distribution parameterisation maths space 36 distribution (field) description maths function XML physical quantity scale: units coordinate system maths space HDF 5 36

Structured data block gigabytes of structured data extract and download subsets variables positions states/times semantics acknowledgements to 37 37

External reference to HDF 5 <Physical. Distribution ID=“my. Distribution”> <sub. Property. Of resource=”. . . #measured. Temperature"/> the property <domain resource="#my. Sensor. Set. At 10. 30"/> the object that has it <range resource=”. . . #Thermodynamic. Temperature"/> <composition. Of> <Physical. Distribution. Description. Vector> <has. Terms parse. Type=“Collection”> <Inverse. Parameterisation>. . . <Inverse. Parameterisation> the data order <Symbol. Table about=“. . . /myfile. hdf 5”/> thedata <Inverse. Scale>. . . </Inverse. Scale> the units </has. Terms> </Physical. Distribution. Description. Vector> </composition. Of> 38 </Physical. Distribution> 38

Conclusions http: //www. scadaonweb. com

Scada. On. Web Ontologies OWL (with RDF and RDFS) W 3 C physical object structure number and text ISO 15926 maths physical property quantity and scale decomposition and sampling physical quantity spaces distribution and description SI units time IEC 61850 quality SCADA 40 demonstrator ontologies properties (ISO 15926) 40

Next steps Automated decision making ð Now – predefined alert (e-mail or txt message) template implemented by explicit code ð Future – alert criteria defined by a rule base Make the semantic measurement object stick as an industry standard ð Publication of measurement ontologies as annexes to existing ISO/IEC standards ð Process industry – links to ISO 15926 (Life cycle data for process plant) ð Electricity transmission grid control standards ð National energy management standards ð Open source semantic measurement software 41 41

Flood warning demonstrator prototype http: //www. scadaonweb. com/demonstrators. html 42 42

Questions http: //www. scadaonweb. com