CAUSALITY AND ENERGY. Prof. A. Bouscayrol (University Lille1, L2EP, MEGEVH, France) Prof. C. C. Chan (University of Hong-Kong, China)

Size: px

Start display at page:

Download "CAUSALITY AND ENERGY. Prof. A. Bouscayrol (University Lille1, L2EP, MEGEVH, France) Prof. C. C. Chan (University of Hong-Kong, China)"

Piers Long
5 years ago
Views:

1 Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» «SYSTEM, CAUSALITY AND ENERGY» Prof. A. Bouscayrol (University Lille1, L2EP, MEGEVH, France) Prof. C. C. Chan (University of Hong-Kong, China) Aalto University based on the Keynote at EMR 09

General Chair of IEEE-VPPC 2010, Lille France Prof. C.C. Chan Tne University of Hong-Kong, China Fellow, Royal Academy of Engineering, U.

2 - Speaker and contributor - 2 Prof. Alain BOUSCAYROL University of Lille 1, L2EP, France Coordinator of MEGEVH, French network on HEVs General Chair of IEEE-VPPC 2010, Lille France Prof. C.C. Chan Tne University of Hong-Kong, China Fellow, Royal Academy of Engineering, U.K. Academician, Chinese Academy of Engineering President, Electric Vehicle Association of Asia Pacific Honorary Professor, University of Hong Kong

3 - Outline - 1. Model, Representation and simulation Different models Different representations Different simulation approaches 3 2. Energy and Systems Systemic approach Energetic Approach 3. Graphical description for engineering Different graphical descriptions Model, description and control 4. Energy management of EVs and HEVs

4 - Philosophy engineering - 4 Six Principles of Integrated System Design Debate, define, revise and pursue the purpose/objective The system exists to deliver capability, the end justifies the means. The statement of a requirement must define how it is to be tested. Requirements reflect the constraints of technology & budgets. Think holistic The whole is more than the sum of the parts and each part is more than a fraction of the whole Be creative See the wood before the trees core of the lecture Follow a disciplined procedure Divide and conquer, combine and rule Take account of the people To err is human ; Ergonomics; Ethics & Trust Manage the project and the relationships All for one, one for all

5 - Energy and Systems: basic requirements - 5 Systems for energy conversion devices in Cybernetics dynamic interactions, organised to achieve a goal Energy nodes interaction principle: Electrical key of management Engineering action and reaction holistic principle: no energy disruption properties induced by associations causality principle multi-finality principle: physical causality is integral several solutions to achieve the objective subjectivity principle: study depending of the user

6 Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» 1. «Model, representation and simulation» Aalto University

7 - Model and representation - Model = description based on physical laws (validity range function of assumptions) 7 Representation = organisation of a model in order to highlight some properties Example state space representation d v 1 dt C c i c C i c v C d i C dt c v c mathematical model transfer function bloc diagram COG i c i c V I c 1 c ( s ) ( s ) 1 Cs Cs v C v C

8 - Simulation of a system (1) - 8 real system model objective assumptions organization no assumption prediction assumptions system model system representation system simulation limited validity range valuable properties behavior study

9 - Simulation of a system (2) - Classical way (e.g. Matlab-Simulink ) 9 real system assumptions no assumption assumptions system model system representation system simulation Intermediary steps are required for complex systems

10 Objectives: component design/optimization component control system analysis (efficiency ) energy management of the system. - Model objectives - different kinds of objectives different kinds of modelling Which model? 10 [Chan 2010] Modelling: structural/functional models static/dynamic models causal/ acausal representations backward/forward simulation...

Design application Functional description function priority Virtual links between subsystems

11 - Structural vs. functional descriptions - How to describe a system? 11 Structural description Physical structure in priority Physical links between subsystems Design application Functional description function priority Virtual links between subsystems Analysis and control application Example v 2 i 1 m v 1 m i 2 3D Finite Element Model Mathematic model Assumption: Ideal transformer

12 - Dedicated software - two DC machine system 12 PSIM (structural) Matlab-Simulink (functionnal) machines connected by a unique link (shaft) machines connected by two links (torque/speed)

13 - Static vs. dynamic models - Which model subsystem? 13 Static model steady state operations no transient states fast computation time global behavior Dynamic model transient state operations but also steady state operations long computation time detailed behavior Quasi-static model static model + main time constant intermediary computation time intermediary behavior

14 Torque in Nm static efficiency map i DC Speed in rpm DC «System, Causality and Energy» 85 Pt (, ) U Example of electrical machine - quasi-static model 85 + J d dt T em T load f T V Sd R S i Sd d Sd S Sq dt V Sq R S i Sq d Sq dt S Sd 0 R R i Rd d Rd dt R Rq 0 R R i Rq d Rq dt R Rd Lm p LR d J T dt em dynamic model 14 ( Rd isq Rq isd em T load f )

15 - Causal vs. non-causal representation - How to connect subsystem? 15 Causal description fixed input and output output = integral function of inputs difficult interconnection subsystems basic solver T 1 T 1 T 2 J d dt T 1 T 2 Non-causal (acausal) description non-fixed inputs and outputs different relationships easy subsystem interconnection specific solver required simulation library T 1 T 2 T 2

16 Example ICE J d dt - Subsystem interconnections - T 1 T 2 causal description 1 T1 T2 T 2 T 3 J d dt 2 T2 T3 electrical machine acausal description 16 T 1 T 2 J 1 J 2 T 1 T 3 J 1 J T 2 1 T 2 T 3 T 2 d ( J1 J 2) T1 T dt 3 derivative relationship T 1 T 3 J equ specific solver

17 - Forward vs. backward simulation - 17 Which method to compute the model? Forward approach from the cause to the effect respect of the energy flow controller required M d dt v F tract F res Backward from the desired effect to the required cause anticipate energy flow no controller required F tract F res v F tract F res v drive cycle F tract-ref control v ref drive cycle derivative relationship (no real-time application)

18 - Example of fuel consumption of a vehicle - 18 consumption d fuel T tract F tract Fuel ICE TM Vehicle v Forward p v F res control v ref drive cycle consumption Backward d fuel T tract F tract Fuel ICE TM Vehicle p v v ref F res drive cycle could be same models, but different representations (cf. I/O)

19 Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» 2. «Energy and Systems» Aalto University

20 HEVs = multi-physical systems multi-layers systems energy nodes. «System, Causality and Energy» - Necessity of a system approach - 20 association of various subsystem In order to combine their advantages Necessity of optimization of cost, management, integration, reliability System is a key word! Objective : 1+ 1 > 2 (Philosophy of engineering) Prof. CC Chan [Chan & al 09]

21 - Systemic approach - System = interconnected subsystems organized for a common objective, in interaction with its environment 21 Systemic = science of study of systems and their interactions holistic property: the system is a whole which cannot be deduced by the study of its subsystems Cartesian approach = the study of subsystems is sufficient to know the system behavior Interactions and associations of subsystems will indicate which approach is required

22 - Cybernetic Systemic - or Black box approach: no internal knowledge 22 in out identification test: observation of out(t) from selected in(t) Behavior model: out(t) = f(t) in(t) in out control out ref closed-loop control of out: for uncertainty compensations

23 - Cognitive Systemic - or White box approach: prior internal knowledge 23 Physical laws of system components Knowledge model: out(t) = f(t) in(t) in out in out control out ref control = inversion of model: (closed loop = an inversion way)

24 - Systemic example - DC machine and smoothing inductor i L f r f u i u 2 24 L m r m u 2 i e u u 2 L f di dt u u2 rf i Lm di dt u2 e rmi L f +L m r f +r m u i e di ( L f Lm ) u e ( rf rm dt )i Association of both subsystems must be studied globally L r f f L r m m L r f f L r m m

25 - Necessity of energetic approach - 25 HEVs = multiple subsystems multiple energy sources energy nodes. Necessity of optimization of efficiency Energy management is a priority Energy accumulator must be carefully manipulated to avoid damages! Energy is a key word! reduction of energy consumption and pollutant emissions

26 - Energetic approach - Energy = amount of work that can be performed by a force, an object a system 26 Ideal energy conversion: energy conservation (no losses) and instantaneous transfer (no delay) but Energy dissipation: losses, reduction of efficiency Energy accumulation: delay in energy transfer Energy accumulation in subsystems is key transformation for safety and efficiency

27 - Energetic examples - 27 C i c d ic C vc v C dt Fast charge: to be controlled to avoid damage Low time constant: few influence on the energy transfer i Bat v Bat Long charge: to be anticipated Long time constant: to be considered in the energy management

28 - Input and output of a system - 28 Input: produced by environment, imposed to the system for evolution (independent of the system) Output: consequence of the system evolution, imposed to its environment (not directly dependant on the environment) Input System Output Environment & System must be defined first! Environment

29 - Interaction principle - 29 Interaction principle Each action induces a reaction S2 action reaction S2 power Example Power exchanged by S1and S2 = action x réaction V bat battery V bat load V bat i load i load battery load P=V bat i load

30 - Causality principle - 30 Principle of causality physical causality is integral input output x? cause effect t xdt OK in real-time area knowledge of past evolution t 1 slope impossible in real-time knowledge of future evolution dx dt

31 Example c c C i c dt v C i C d - Causality principle - v 2 c v c E 1 2 i c v C delay no energy disruption 31 v C d dt i c For energetic systems physical causality is VITAL risk of damage

32 Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» 3. «Graphical descriptions for system engineering» Aalto University

33 - Graphical description - Use of GRAPHICAL DESCRIPTIONS for modelling and control of non-elementary systems 33 intermediary step for another view of the system synthetic description respect of physic properties linked to classical modelling transportation systems renewable energy applications production machines drives in industry process tactile interfaces... new ways to design, analyse simulate and manage such systems System: sub-systems in interactions organised for a common objective Remember, See the wood before the trees!

34 - Example of a railway traction system - 34 v rame + i He1 u i He1 ee1 voiture C M1 F res u filtr i hach e V DC ihi u Hi i filtre i ei B1 C M2 B2 u He2 i ee2 causality? action/reaction? V DC k s i filtrre i hach k s i He1 u filtrre c He1 c Hi c He2 [K] [K] i He [K] i He2 i He2 x x x x x x u He1 u He2 k s u He i ee1 k mcc k s k 2 i ee s k mcc Simplified block diagram i ei e ee1 eee2 x x x x B1 C M1 C M1 k bog k bog k bog k bog F bog1 F tot F bog2 F res M s v rame

35 - Example of a railway traction system - 35 v rame + i He1 u i He1 ee1 voiture C M1 F res causality OK action/reaction? u filtr i hach e V DC ihi u Hi i filtre Causal Ordering Graph (COG) i He2 i ei u He2 i ee2 B1 C M2 B2 [Hautier 96] [Guillaud 01] u Hi c He1 B1 v rame V DC i He1 i filtrre u filtrre u filtrre i ee1 u Hi i ei i ee1 i ee1 C M1 F bog1 v rame i He2 c Hi e ei i Hi u filtrre C M2 F bog2 F res u He2 i ee2 B2 c He2 i He2

36 - Example of a railway traction system - 36 v rame + i He1 u i He1 ee1 voiture C M1 F res Bond Graph (BG) u filtr i hach e V DC ihi u Hi i filtre physical causality? action/reaction OK i He2 R : R ex1 i ei u He2 i ee2 B1 C M2 B2 [Paynter 61] [Dauphin 99] R : R f c He1 u filtre i He1 MTF u He1 i ee1 1 I : L ex1 R : Rind1 R : R ind2 e ei1 i ei MGY C M1 B1 TF Fbog1 v rame Se : V DC V DC u filtre i filtre 1 i filtre 0 u filtre i Hi u Hi i ei MTF 1 1 I : M v rame 1 F res v rame Charge I : L f u filtre C : C f i He2 c He1 c Hi u He2 MTF 1 i ee2 R : R ex1 I : L ind1 I : L ind2 e ei2 i ei MGY C M2 B2 F bog2 TF vrame I : L ex1

37 physical causality OK action/reaction OK + - Example of a railway traction system - i He1 u i He1 ee1 C M1 v rame voiture F res 37 u filtr i hach e V DC ihi u Hi i filtre Energetic Macroscopic Representation (EMR) rail SE V DC filtre i filtre u filtre i hach mise en // u filtre i He1 u filtre i Hi hacheurs He 1 Hi c He1 i He2 enroulements u He1 i ee1 u Hi i ei i ee1 i ei e ei i ei u He2 i ee2 mise en série e ee1 i ei eei1 i ei e ei2 B1 conv. EM M 1 C M1 B1 C M2 B2 bogie B 2 F B1 v rame couplage F tot v rame [Bouscayrol 00] [Bouscayrol 05] rame v rame environnement F res SM u filtre i He2 He 2 c Hi u He2 i ee2 i ee2 e ee2 M 2 C M2 B2 B 2 F B2 v rame c He2

38 - Comparison of modelling tools - Energy & System 38 Energetic Puzzles (Laplace, France) Bond Graph (USA, The Netherlands ) Power Oriented Graph (Italy) Signal Flow Diagram (Germany, Japan...)... analysis design 1 0 mathematical model global controls COG (L2EP-LEEI, France) EMR (L2EP, France) causal descriptions for simulation and control Remember, divide and conquer! inversion graphs cascaded control

39 Bond Graph system analysis and design (structural approach) «System, Causality and Energy» - Graphical modelling tools - use of Petri Nets discrete event systems Causal Ordering Graph (COG) drive control (functional approach) 39 Energetic Macroscopic Representation (EMR) system control (functional approach) Mechanical Engineering Power Electronics Electric Drives France Electrical Engineering USA The Netherlands worldwide Electric Systems France Energetic Systems Canada Switzerland Denmark China

40 Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» 4. «Application to energy management of EVs and HEVs» Aalto University

41 «System, Causality and -- Energy» - Which model for EV/HEV control? - 41 Multi-physical system Systemic approach Energy management System control Real-time control Energetic approach Causal modeling Functional description Dynamical modeling Causal modeling Moreover a graphical description could be a valuable intermediary step for such complex systems

42 - Different control levels - 42 Energy management of HEVs: Energy management of local subsystems Energy management of the whole system (co-ordination of subsystems) Two control levels can be organized: - local control - system supervision Quasi-static models Dynamic and causal models [Delarue & al 2005] compatibility of the control levels compatibility in term of inputs/outputs

43 Parallel HEV «System, Causality and Energy» - Different control levels (example) - BAT VSI Fuel EM ICE Trans. 43 fast subsystem controls EM control ICE control Trans control slow system supervision Energy management (supervision/strategy) driver request

44 - Local control - 44 Local energy management: must take into account power flows in all parts of subsystems EM control ICE control Trans control Classical controls of subsystems: required dynamic and energetic models to manage power flows in real-time

45 Rule-based «System, Causality and Energy» - Energy management methods - deterministic rule-based fuzzy rule-based state machine / power follower/ thermostat control predictive / adpative / conventional 45 Optimization based global optimization real-time optimization Dynamic programming / stochastic DP / Game theory / Optimal control. Robust control / Model predictive / decoupling control / l control Energy management (supervision/strategy) [Salmasi 2007] driver request

Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» «Conclusion» system = subsystems in interaction best performances require a systemic

46 Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» «Conclusion» system = subsystems in interaction best performances require a systemic approach energy = respect of the physical causality energy management requires a causal approach control -> inversion of a causal model of the system in order to respect its energy properties graphical description = model organization useful intermediary step Aalto University Remember, follow a disciplined procedure!

47 - References - P. J. Barre, & al, "Inversion-based control of electromechanical systems using causal graphical descriptions", IEEE-IECON'06, Paris, November A. Bouscayrol, & al. "Multimachine Multiconverter System: application for electromechanical drives", European Physics Journal - Applied Physics, vol. 10, no. 2, May 2000, pp (common paper GREEN Nancy, L2EP Lille and LEEI Toulouse, according to the SMM project of the GDR-SDSE). A. Bouscayrol, G. Dauphin-Tanguy, R Schoenfeld, A. Pennamen, X. Guillaud, G.-H. Geitner, "Different energetic descriptions for electromechanical systems", EPE'05, Dresden (Germany), September (common paper of L2EP, LAGIS and University Dresden). C.C. Chan, The state of the art of electric, hybrid, and fuel cell vehicles", Proceedings of the IEEE, Vol. 95, No.4, pp , April C.C. Chan, A. Bouscayrol, K. Chen, "Philosophy of Engineering and Modelling of Electric Drives, International Conference on Electrical, Keynote, October 2008, Wuhan (China) C. C. Chan, A. Bouscayrol, K. Chen, Electric, Hybrid and Fuel Cell Vehicles: Architectures and Modeling", IEEE transactions on Vehicular Technology, vol. 59, no. 2, February 2010, pp (common paper of L2EP Lille and Honk-Kong University). G. H. Geitner, "Power Flow Diagrams Using a Bond graph Library under Simulink", IEEE-IECON'06, Paris, November J. P. Hautier, P. J. Barre, "The causal ordering graph - A tool for modelling and control law synthesis", Studies in Informatics and Control Journal, vol. 13, no. 4, December 2004, pp H. Paynter, "Analysis and design of engineering systems", MIT Press, F. R. Salmasi, "Control strategies for Hybrid Electric Vehicles: evolution, classification, comparison and future trends", IEEE Trans. on Vehicular Technology, September 2007, Vol. 56, No. 3, pp R. Zanasi, R. Morselli, "Modeling of Automotive Control Systems Using Power Oriented Graphs", IEEE-IECON'06, Paris, November 2006.

48 Aalto University Finland May 2011 «Energy Management of EVs & HEVs using Energetic Macroscopic Representation» «SYSTEM, CAUSALITY AND ENERGY» Prof. A. Bouscayrol (University Lille1, L2EP, MEGEVH, France) Prof. C. C. Chan (University of Hong-Kong, China) Aalto University based on the Keynote at EMR 09

«SYSTEM, ENERGY AND CAUSALITY»

EMR 17 Lille June 2017 Summer School EMR 17 Energetic Macroscopic Representation «SYSTEM, ENERGY AND CAUSALITY» Prof. Alain BOUSCAYROL (L2EP, University Lille1, France) Prof. C.C. CHAN (University of Hong-Kong,