In section 3 the many advantages of sodium cooled reactors were enumerated in comparison with light water reactors. Mainly, these advantages follow from the relatively low energy content of the primary coolant. LWRs are required to deal with the prospect of a primary piping system failure which pressurizes and heats the containment building and requires containment cooling as well as both high and low pressure water makeup sources to ensure the core is kept covered with water. These various LWR safety systems require that there be reliable electric power in large amounts to power very large motors driving the vital safety system pumps.
The governing rule is Criterion 17 from 10CFR50 Appendix A and is stated below:
Criterion 17—Electric power systems. An onsite electric power system and an offsite electric power system shall be provided to permit functioning of structures, systems, and components important to safety. The safety function for each system (assuming the other system is not functioning) shall be to provide sufficient capacity and capability to assure that (1) specified acceptable fuel design limits and design conditions of the reactor coolant pressure boundary are not exceeded as a result of anticipated operational occurrences and (2) the core is cooled and containment integrity and other vital functions are maintained in the event of postulated accidents.
The onsite electric power supplies, including the batteries, and the onsite electric distribution system, shall have sufficient independence, redundancy, and testability to perform their safety functions assuming a single failure.
Electric power from the transmission network to the onsite electric distribution system shall be supplied by two physically independent circuits (not necessarily on separate rights of way) designed and located so as to minimize to the extent practical the likelihood of their simultaneous failure under operating and postulated accident and environmental conditions. A switchyard common to both circuits is acceptable. Each of these circuits shall be designed to be available in sufficient time following a loss of all onsite alternating current power supplies and the other offsite electric power circuit, to assure that specified acceptable fuel design limits and design conditions of the reactor coolant pressure boundary are not exceeded. One of these circuits shall be designed to be available within a few seconds following a loss-of-coolant accident to assure that core cooling, containment integrity, and other vital safety functions are maintained.
Provisions shall be included to minimize the probability of losing electric power from any of the remaining supplies as a result of, or coincident with, the loss of power generated by the nuclear power unit, the loss of power from the transmission network, or the loss of power from the onsite electric power supplies.
This criterion or something very close to it governed the design of the electric power system at Fukushima. Obviously, the criterion is not fool proof. It would be much better for nuclear power plants to be able to accommodate total loss of electric power from all sources without experiencing serious consequences. The design approach being presented has that capability. It is useful to consider the 1E loads on the CRBRP and evaluate what has been eliminated using the approach of this paper.
The CRBRP design provided for three separate diesel generators. The underlying reason for three was related to the choice of the SGAHRS for decay heat removal. The thinking was that if one loop were disabled, the single failure criterion required three separate diesel generators supplying three class 1E “divisions”. Divisions 1 and 2 supplied essentially identical loads in their respective divisions while division 3 was provided primarily to supply loop 3 SGAHRS. The CRBRP project had not selected a diesel generator at the time of the project’s termination but the peak loads on diesels 1 and 2 including a 15% allowance for expansion were expected to be around 3400 KW. This is equivalent to about 4600 horsepower – very stout diesels indeed. At these large sizes, diesels become difficult to start and require extensive maintenance and frequent testing to assure reasonable reliability. If a diesel is taken out of service during plant operation, operational constraints are likely to be applied by technical specifications because of the single failure criterion. If the engines could be much smaller some of these issues would vanish.
In order to evaluate possibilities, again CRBRP will be used as a convenient point of departure in view of the abundant information available in the PSAR. As stated above the 1E loads were estimated to be 3400 KW for each of divisions 1 & 2, which were based on an identified load of 2967 KW. These loads can be conveniently divided into eleven groups which will be separately treated below.
- Steam Generator Auxiliary Heat Removal (SGAHRS) 1414.7KW By far the largest emergency load is that associated with the SGAHRS and their associated auxiliary feed water pumps which is replaced with naturally circulating systems in the proposed design that require no power other than instrumentation.
- Ex-Vessel Storage Tank (EVST) 253 KW The proposed design approach would replace the active system used in CRBRP with a passively cooled system. As above, the only load remaining would be for monitoring instrumentation.
- Annulus cooling 1268.2 KW The annulus cooling system was part of the CRBRP system for dealing with the HCDA, which is considered inappropriate as a design basis. Of this total load, 1103 KW occurs when there is no SGAHRS load – presumably because the core is assumed to be on the floor of the reactor vault so SGAHRS wouldn’t do much good. The remainder, 165.2 KW occurs coincident with SGAHRS operation and will be considered here as a deduction.
- Diesel support 29.6 KW This number includes such things as fuel supply and engine cooling. It probably cannot be eliminated altogether but likely can be reduced with smaller engines.
- Control Room support 81.2 KW This includes supply, return, and filter fans necessary to maintain an habitable environment in the control room and can probably not be eliminated. However, modern control rooms will inevitably be much smaller with much less electric power loads than was the case for CRBRP, which was patterned after nuclear plant control rooms that were extant at the time.
- Containment isolation 4.8 KW This is a small load provided to supply power to valves necessary to isolate the containment that probably cannot be avoided.
- Primary sodium make-up pump 18KW This pump was required to be supplied with 1E power because of the DHRS employed on CRBRP. Since the proposed design includes this feature, it is retained as an emergency load.
- Lighting 50 KW This is lighting for areas of the plant necessary to be illuminated to enable operators to accomplish and maintain safe shutdown. A number similar to this will likely apply to any nuclear plant although it can probably be reduced by using the more efficient systems available today combined with the reduced size control room.
- Battery charger 115.5 KW It is not clear why the CRBRP designers used the full load of the battery chargers to prepare their load list. The battery would be fully charged whenever there was a loss of off-site power so all that would be required would be enough power to supply the DC load. On CRBRP, the DC load was about 50 KW per division, which will be used for this analysis. There is no reason to expect it to scale with plant size.
- Emergency chilled water (including chiller) 583.5 KW Slightly more than half of the CRBRP emergency chilled water load is for systems such as SGAHRS and EVST cooling that is not applicable to the concept proposed here. The control room load is very high and can certainly be reduced for a more modern control room design with fewer heat sources.
- Emergency service water 176.4 KW This system cools the diesels, provides the heat sink for the emergency chilled water system and supplies the fire protection system. The system is cooled by a tower that requires power for its fans. With smaller diesels and a smaller chilled water system, this load can probably be reduced somewhat.
The table below summarizes and compares CRBRP requirements with the design approach proposed and constitutes CRM 42:
|3. Containment annulus||165.22||0|
|4. Diesel support||29.6||15|
|5. Control room support||81.2||50|
|6. Containment isolation||4.8||5|
|7. Primary sodium makeup pump||18||40|
|9. Battery charger||115.5||50|
|10. Emergency chilled water||583.5||200|
|11. Emergency service water||176.4||90|
Table 3 Emergency loads per division, KW
Allowing 15% growth in loads as was done in the CRBRP PSAR leads to the requirement for 700 horsepower diesels and there would be just two of them. At this much smaller size, many options present themselves for diesels that are much more reliable, easier to start, and cheaper than the three 4600 horsepower machines that would have been required on CRBRP. This is also much smaller than the emergency diesels required on LWR plants offering another competitive advantage for the LMFBR.
However, there is an important distinction in the above table between the emergency loads for CRBRP and those for the proposed design. For the case of CRBRP, the loads are those necessary to maintain the plant in a safe state, e.g. decay heat removal from both the reactor and the EVST. (This is somewhat fictitious since it was shown in Section 8 that CRBRP actually had station blackout capability, at least for the decay heat removal system if not for the EVST. Despite this capability, as was stated in Section 8, the project never sought nor realized any licensing benefit from the capability.) For the proposed design, electric power necessary to maintain safe shutdown has been reduced to zero and the loads defined are those that would be desirable to be supplied in the event of loss of all offsite power. This constitutes CRM 43. It would be desirable to maintain control room habitability and lighting but not essential for safe shutdown. In the highly unusual event of a station blackout it may be necessary to monitor instrumentation locally with some sort of battery operated portable power supply and there likely would be the occasion for checking the status of valves by viewing manual indicators, or throttling back on naturally circulating air cooling of decay heat removal systems but a committed 1E power source is not needed. Thus, the emergency diesels provided to accommodate loss of offsite power and the entire emergency power supply may not need to be 1E.
A careful definition of how this would be done would be a worthy project for further study. Operating a plant that is designed for station blackout and knowledge of exactly what is to be done by the operators of the plant should one occur is a far better alternative than being reliant on large 1E diesel generators and experiencing core damage when a station blackout occurs.
There is one load on the list above that warrants special attention – the primary sodium makeup pumps. As the reactor cools down, the sodium level in the reactor vessel will drop. It is necessary to maintain the outlet nozzles covered to ensure primary system flow to the PRACS heat exchangers in the IHXs. There are four possible approaches for dealing with this issue. 1) A DRACS could be incorporated into the reactor vessel. With open vessel refueling, a DRACS would require its own nozzles to carry the NaK to the DRACS cooler and could complicate refueling. With the single rotating plug concept, DRACS would be entirely feasible, penetrating the horizontal baffle and head outside the rotating plug. The DRACS heat exchanger(s) would need to be located adjacent to the vessel wall in the outlet plenum so as not to interfere with refueling and there would need to be one or two down-comers within the vessel penetrating the core support cone to return the cooled sodium to the inlet plenum. 2) The reactor vessel height could be increased so there would be no possibility of uncovering the outlet nozzles. Using a volumetric coefficient of expansion of sodium of 1.6 X 10-4/°F, accommodation of 10% volume change from 400°F to 1000°F could be accomplished by adding two ft. to the reactor length. 3) A 1E power supply expressly committed to the makeup pumps could be adopted. Since the makeup pumps are small loads, the power supply (about 40 KW) could be furnished by batteries, but the pumps must be capable of continuous operation if they are necessary for OHRS function. 4) Since the overflow tank is provided with a cover gas, overflow tank inventory could be transferred to the reactor vessel by pressurizing the overflow tank above reactor cover gas pressure.
1 Figures are drawn from Table 8.3-1A in the CRBRP PSAR.
2 The annulus loads following a postulated HCDA are predicted to be 1268.2 KW.
3 Note that the numbers in this column do not add to this figure despite the fact that they were all drawn from the same table in the PSAR. The number shown is the figure cited in the PSAR table as the total.