for ELFORSK (the Electrical Utilities R & D Company, Sweden)

Transcription

1 Economic aspects on Flexible Production and the Nordic Market IAEA-TM on Flexible Operation Approaches for Nuclear Power Plants 4-6 Sept 2013 Hans Henriksson, Jonas Persson Vattenfall R&D, Sweden for ELFORSK (the Electrical Utilities R & D Company, Sweden)

2 Introduction Future requirements could demand a more flexible generation at our Nuclear Power Plants (NPPs): The variation in production and consumer load will be much larger and faster than before. New electric grid code from ENTSO-E with tougher regulations. A flexible power control of NPPs could, with increased intermittent energy production, be a valuable asset. 2

3 Background What is flexible power generation? In Sweden nuclear power is seen as base load power: 100% constant power. Internationally three different modes of operation have been used: Primary: automatic frequency control, (seconds) Secondary: electric spot price (minutes-hours) Load-following: power follows demand: (e.g. 100% during the day, 50% at night or weekend) Why load-following? More intermittent power in the grid New grid requirements (ENTSO-E) New market (four regions in Sweden) New transmission links to other markets 3

4 Previous studies on flexible nuclear power One study, carried out in 2011 covered the possibility to load-follow with nuclear power: Elforsk 12:08 (in Swedish) A follow-up report focused on additional costs due to load-following, Elforsk report 12:71 Conclusions: Load-following would mean only minor additional operational costs. If planned properly, these costs may even be avoided. Fuel cost increase (up to 4% increase of production price in a BWR). Slightly higher relative fuel cost in PWRs. Additional maintenance in control rod drive mechanism decreased availability (<1,8%) in France (due to primary control operation). It is recommended to use several reactors in series to avoid larger power decreases, below 60% nominal reactor power. 4

5 Overview of study A study on additional costs due to load-following nuclear power was carried out in 2012 focusing on: Component wear and tear Operation Maintenance Fuel costs Experience from Sweden, Finland, France, Germany: Interviews with FKA, RAB, OKG from previous experience Fuel load modelling for BWR and PWR cycles Visit to EnBW and Philippsburg I, Sept: Roman Zofka, Jörg Storbeck et c. Visit to EDF and Nogent-sur-Seine, Nov: Ph. Lebreton acknowledged 5

6 Limited experience of load-following operation in Sweden Weekend Nights 6

7 Limited experience of load-following operation in Sweden Swedish Nuclear Power Plants are co-owned by several utilities. If one owner demands a power decrease, compensation per MW is granted to the other owners due to possible wear and tear. The main reactor type is the BWR. These are regulated by main re-circulation pumps (to 60-70% power). Forsmark was for example equipped with Frequency control units originally, but this was removed in the early 1980s. 7

8 Experience from Finland Limited use Load-following is prohibited according to the technical specifications and safety rules (in Finnish TTKE) if not a special permit has been granted by STUK (the safety authority in Finland) In 1981 IVO (former name of Fortum) investigated the possibility of loadfollowing. The obtained permit was granted with the conditions that: Maximum 3 unrestricted regulations every year. Maximum 7 regulations with maximum 100 MW per plant every year. This kind of operation has been used (data from 2001 shown), but is limited to some weekends or holidays. The operation is limited by boric acid control, as the control rod positions are strongly regulated by STUK and in principle control rods are always withdrawn during operation. 8

9 Experience from Germany (example from Philippsburg) Extensive experience from flexible operation Primary and secondary control has been used extensively Load-following is possible with a notice of 5 h in advance in Germany. Normally this is done by reducing power progressively between NPPs, in this case KKP1 (BWR) and KKP2 (PWR): First the operations at KKP1 decrease to 74% of full power (KKP2 unchanged). Thereafter, KKP2 power is decreased to 70% (KKP1 unchanged) Further decrease is effectuated by KKP1 followed again by KKP2 (see below) No indication of fuel damage has been seen due to load-following 9

10 Experience in France In France, Nuclear Power Plants (NPPs) are used in flexible operation: 1. Primary control (automatic frequency control). 2. Secondary control. 3. Power generation in response to load: Load-following RTE (TSO in France) is in charge for this control and discusses directly with EDF (the utility and NPP owner in France). Less neutron-absorbing grey control rods are used for Primary control. In this way, the (axial) core power distribution can be smoother. The power control is performed from a level of 93% of rated thermal power, leaving a margin of 2% frequency compensation, and 5% margin to secondary control. Load-following is carried out in the power range 37%-93% during the first 85% of the fuel cycle. (Note, the French fleet consists of PWRs only) 10

11 International examples: France After a visit to EDF, France, a few additional costs were identified: The availability in France is 1.8% lower due to flexible generation (mainly due to primary control) Wear on control rod driver mechanisms causes an increase in maintenance 11

12 Fuel costs: Modelling 1+5 fuel cycles A study of additional fuel costs was performed as well. Four scenarios with unplanned load-following were considered (six complete fuel cycles with the first cycle modified) 1. Reference fuel cycle (including coast-down), 2. the first cycle as reference cycle with 20% shorter cycle length, 3. the first cycle operated at 60% power first half, 4. the first cycle operated at 60% power second half. The need of fresh fuel assemblies (FA) in the reference case is 742, while for the other three However the power ratio is EFPH/FA for case 1, while it is about 72.5 EFPH/FA, (i.e., less energy output per fuel assembly). No difference between cases 3 and 4 indicates limited effect of spectral shift Fuel costs (about 20% of total production cost): The additional fuel cost for BWRs could be 17-23%. For PWRs, the additional cost could be %. 12

13 Fuel costs: Modelling 1+5 fuel cycles A rule of thumb states that cost of surplus of fresh FAs compared to fuel demand may be calculated by dividing the cost of extra FAs with the system number (number of cycles the FA spend in the core for equilibrium conditions). Reactor: BWR, Forsmark 1 PWR, Ringhals 4 Scenario 2 (20% shorter cycle length) Scenario 3/4 (operation at 60% power during ½ cycle) Scenario 2 (20% shorter cycle length) Cost increase, % (rule of thumb) Cost increase, % (in this study) Difference Fuel costs (which accounts for <20% of total production cost): The additional fuel cost for BWRs could be 17-23%. For PWRs, the additional cost could be %. 13

14 Results (2) The optimal working point cannot always be reached, i.e. lower turbine efficiency and less efficient Pu-burn. The power increase is hampered by new thermal equilibrium of fuel, xenon-poisoning, et c. It is recommended to use several reactors in series to avoid larger power decreases, and hence worsened working points for pumps. The optimal (and less demanding) power range is % of rated thermal power. 14

15 Results: additional costs during load-following Only minor additional operational costs. If planned properly, these costs may even be avoided. Additional maintenance in control rod drive mechanism decrease availability (<1,8%) (seen in France during primary control operation). No load-following is carried out in PWRs after 85% of the fuel cycle due to a limited reactivity margin, and in the case of fuel failure in the core. The additional fuel cost for BWRs could be 17-23%. For PWRs, the additional cost could be %. 15

16 Risks Damaged fuel Limits in operation and flexibility when having damaged fuel Under-loading fuel Would affect fuel cycle length and/or coast-down period. With strong seasonal variations in the Nordic countries, outage is preferable in the summer seasion, hence 11 month fuel cycles Component wear The power ramps need to be performed according to specifications to avoid wear. As long as no design transient specifications (DTS) are used no risk of additional wear is seen Increased risk of operational disturbancies Every change to steady-state operation could case disturbancy No indications of such situation has been given. 16

17 Outlook (1) How precise can power production variations be in present power plants? How would primary and secondary control be carried out, as compared to load-following in a nuclear power plant? What are the properties of the plant control loop when production is varied, i.e. what is the time constant (delay)? What are the manpower and competence needs for flexible power production? Possible constraints in operation during flexible power generation? (maintenance, safety, ) 17

18 Outlook (2) Another option would be to consider the thermal power output in a more flexible way (operating with constant reactor thermal power) The balance between generated electric power and for instance district heating could be more flexible. Is this only hypothetical for old/present reactors? Is it feasible for new plants? Investigate specific needs for BWRs, PWRs Could a heat storage unit be coupled to the steam cycle for the optimised power output? 18

19 Outlook (3) Possible effects on the power system, restrictions from normal behaviour: Examination of the NPP s possibility to deliver reactive power to the grid (voltage stability) and if this is affected by any mode of operation. Can the plant still minimize system oscillations with the help of a Power System Stabilizer (PSS) during these periods? Are there other properties that occur with respect to the grid? 19

20 Price areas in Sweden A large new intermittent power production, i.e., wind power that influence the power system and the market. Intermittent power production demands more regulating power. Therefore, can NPPs also provide flexible power production to the market such as primary and secondary control? Price area 1 Price area 2 Price area 3 Price area 4 20

21 Expansion of wind power in Sweden The production from wind power 2010 was 3.5 TWh. The production from wind power 2013 will be ~9.3 TWh. The production from wind power 2014 will be ~10.7 TWh. At January 1, 2014, Sweden will have 4.3 GW of installed wind power. 21

22 Variations Wind power has a variable production and the production is also difiicult to predict Even though the wind power is fairly good distributed along Sweden, its production varies between 50 MW to 1800 MW (2011) : : : : : : :00 The consumption in Sweden varies during a normal day with 5000 MW. When the wind power will reach the installed capacity of MW (2030), then its variation will be larger than the variation of the consumption. 22

23 The market More wind power will lead to higher volatility on the prices (Elspot and regulating prices). This means that the ancillary services below can be of more interest to deliver. Frequency control Primary control Secondary control 23

24 Frequency control The Swedish TSO (Svenska Kraftnät) has the responsibility to keep the power production in balance. The used power for this is ~2,5 TWh per year which is not so much compared to the total consumption of 150 TWh. Svenska Kraftnät procures the primary control and from April 1, 2011 it is split in two products Frequency Controlled Normal Operation Reserve (FNR) and Frequency Controlled Disturbance Reserve. FNR is used during normal operation (49.9 f 50.1 [Hz]). FDR is used during disturbances, i.e., when the frequency goes below 49.9 Hz ( 49.5 f < 49.9 [Hz]). 98% of Vattenfall s regulating power is situated in Price areas 1 and 2. FDR is ~400 MW in Sweden. 24

25 Secondary control The secondary control in the Nordic synchronous power system is done by bids which normally are sent to Svenska Kraftnät not later than 45 minutes before the operational hour. The response time for a bid can be 1, 5, 10, or 15 minutes. The aim with secondary control is to come back to Hz so that the primary control can recover, i.e., is relieved. The secondary control is today ordered manually. Today a third of the operational hours are unregulated, the rest of the hours are up or down regulated. With more wind power we expect that more hours have to be regulated and also larger volumes that can increase the regulating prices. 25

26 An hour of a normal Nordel day The frequency varies between < f < Hz at July 2, The frequency is fairly uncontrolled and sometimes outside the normal band width of 49.9 < f < 50.1 Hz. 26

27 Secondary control (LFC tests) During January April, 2013 it has been tested an automatic secondary control in the Nordic Power System. During one week 400 MW of Load Frequency Control (LFC) was applied. See the next slide. 27

28 A Nordel day with 400 MW LFC The frequency varies between < f < Hz between March 18, The secondary control seems to work and is turning the frequency back to 50,00 Hz. 28

29 Thank you! 29