Mitigating power fluctuations on utility grid created by klystron modulators

1. Mitigating power fluctuations on utility grid created by klystron modulators Sklavounou Eleni TE – EPC – FPC

2. Introduction • Klystron Modulators R&D Objectives: • Propose design solution for the modulators charging sub-system minimizing the grid power fluctuation. • Design the control strategy of capacitor chargers for minimum power fluctuation. • Problematic: Grid power fluctuation VS capacitors voltage droop + current and voltage regulation capabilities of the charger + cost + size

3. Industrial survey for chargers and capacitors [1] • Contacted companies for capacitor charger offers: Fug, Technix, Heinzinger. • Offers received for constant power consumption from the utility grid: Aiming for 96% minimum charger efficiency

4. Industrial survey for chargers and capacitors [2] • Main capacitor bank: • TYPE: Film capacitor. • SIZE: Create a model describing volume vs stored energy of a capacitor bank. (Data obtained from: AVX Film Capacitors) Voltage level was taken into consideration BUT finally it is depending on Energy only!

5. Power quality specifications • Network specifications according to IEC and EDF • We don’t have all the specifications yet, however we are creating models which allow us to predict power quality vs different design choices. Once the power quality specs are defined, models can quickly provide hints for design directions!

6. Design considerations for power fluctuation mitigation • Considering the charger as a constant current source: Unacceptable power fluctuation!!!

7. Ideal case: Imposing Pch=constant Ideal charging current Very high bandwidth charger needed

8. Real case study: model the capacitor charger as 2nd order transfer function: *Ideal case: infinite current bandwidth *Real case study: 2kHz current bandwidth

9. First results: Power fluctuation of the charger as a function of its closed loop current bandwidth.

10. Get into the charger topology • Internal capacitor bank as a second “filter” to • Stabilize the voltage at the input of the DC/DC converter. • Provide active power filtering to the grid • Which are the compromises between the internal capacitance Cinternal and the DC/DC converter bandwidth?

11. 1st results for the grid active power • Power fluctuation for different Cint and Bandwidth of the DC/DC converter (transfer function modeling approach)

12. Simplified models for power & charging voltage control studies • DC/DC conversion topology : buck converter (benefit from the linearity between duty cycle and output voltage). • We aim to simplify our analyses, especially when considering several systems in parallel. • Verify the accuracy of the created models by comparing the output of the transfer function modeling approach with the output of the equivalent power model, considering the same input.

13. Models simplification • Power model (time domain) • Slow • Too complex (depending on topology) • For overall power & voltage control strategies - Not interested in design details • Transfer function model (frequency domain) • Fast • Simple • Neglecting phenomena at switching frequency level (we are not interested in!) • Handful for control analyses Need to cross-validate the two approaches to verify that simplifications are acceptable.

14. Validation for equivalence between models Equivalence between time domain and frequency domain models

15. Active power closed loop operation problematic vs loading condition changes Estimated averaged active power Unexpected load conditions (i.e. main PFS switch timing deviation) Result???

16. Multi objective control needs • …but the charging voltage should be reproducible! • We need multi objective control strategies (for power fluctuation & voltage reproducibility). • …or one can imagine a correction on the charging current! • We are studying methods to derive multi-objective control strategies for finding the optimal compromise between Vch and Pfluct. For instance

17. Active power measurements & experimental results [1] • Try to implement control strategies in practice with existing charger. • The control strategy could be systematically used in EPC capacitor discharge converters to reduce power fluctuations • Grid current measurements: • Results: Quality of Technix chargers (diode bridge topology, no PFC → insert harmonics in the current → Not suitable for power leveling control implementation).

18. Active power measurements & experimental results [2] • Measurement of 3phase active power by measuring V and I and implementing Clarke transform • Difference between constant charging current values

19. Conclusions & Future work • Conclusions • Power Design: -Survey of commercial solutions – cost, size, regulation capabilities-Correlate charger DC/DC conv. bandwidth with grid power fluctuation and charger size • Control Design:Simplified & validated models for power & charging voltage control studies • Currently evaluating control strategies • Future work • Power Design:Review and design of the converters (performances evaluation of different topologies) • Control Design:Continue control studies to optimise compromises between active power fluctuation & charging voltage reproducibility

20. Ευχαριστώ πολύ Thank you very much Merci beaucoup