Despite quantitative uncertainty, in this section, the generated valid procedures are compared with existing SAM strategies that are primarily described in the U.S. Nuclear Regulatory Commission (NRC) report NUREG/CR-5869 titled "Identification and Assessment of BWR In-Vessel Severe Accident Mitigation Strategies" [107]. Whether existing strategies can be covered by the generated procedures, and most importantly, whether there are unmentioned strategies generated are demonstrated. Recent lessons from the Fukushima disaster [6] will be also used as references for comparisons. In the end, both the merits and limitations of the developed planning technique will be summarized.
Table 6.5, which is summarized from [107], lists potential strategies that can mitigate BWR emergency situations resulted from station blackout. There are eleven strategies without detailed procedures given, among which some involve in advance preparation for resources and equipment. While the other may require human’s capabilities of immediate response. Operators need to take advantages of existing and available equipment, as well as their skills and creativity to find solutions for achieving safety goals. Recommendations for immediate response will be analyzed, which show how available equipment can be applied in flexible and previously unassociated ways. In addition, the results will also give suggestions how resources should be reasonably prepared in advance. Insights obtained from comparison are summarized as follows:
• There are many generated procedures using CST as the injection source (Procedures-4, 8, 11, 13, 15, 17, and 19 in Table 6.3, and Procedures-1, 5, and 7 in Table 6.4). In order to make these multiple strategies available, apparently, it is important to keep capacity
Table 6.5 Existing SAM strategies coping with station blackout
Item Strategy Actions or cautions Planning method
Advance preparation
Immediate response
1 Maintain CST as an injection source
(1) Augment and maintain the original water supply in CST.
(2) Use unusual methods of supplying makeup to CST.
√ √
2 Alternative sources for vessel injection
(1) Consider the other sources, e.g. ocean.
(2) Consider clever methods for the introduction of such sources, e.g. temporary hose connections
√
3
Override of injection pump trip (primarily to maintain RCIC or HPCI available)
(1) Maintain operation of vessel injection beyond the point at which they would normally trip.
√
4 Load-shedding to conserve battery power
(1) Shed nonessential DC loads.
(for consideration of continuous operation of RCIC or HPCI)
√
5 Battery recharging to maintain DC loads
(1) Provide a portable charger.
(2) Provide hookup to the DC power system
√ √
6 Replenish pneumatic supply for air-operated SRVs
(1) Backup supplies of control air for SRVs (depressurization).
√
7 Emergency cross-tie of AC power sources
(1) Provide cross-tie capability between independent AC power sources, such as from multi-unit site.
√ √
8 Alternate power supply for vessel injection of CRD
(1) Use mobile diesel generators, or gas-turbine generator to derive CRD pump.
√ √
9 Use diesel-driven FP pumps
for vessel injection or containment spray
(1) Provide cross-connection with vessel injection or containment spray piping.
√
10 Automation for RPV depressurization
by SRVs Improve reliability of actuation logic √
11 Venting
(1) Improve reliability of relevant equipment
(2) Act when necessary.
√ √
of CST, as indicated byStrategy-1in Table 6.5. Although the current MFM model of BWR does not describe functions beyond the independent water sources, which means that no strategy are generated for actions for CST, operators should take the high priority for supplying CST even by using unusual sources (Strategy-2) during immediate response.
• Strategy-2also mentions to consider clever methods for introduction of some sources.
Procedures-8, 14, 15, 16, 17, 19, and 20 in Table 6.3, and Procedures-4, 5, 6, and 7 in Table 6.4 use different unoriginal ways for injecting water into RPV or PCV by available power. For example of Procedure-14, using RCIC as the power and LPCI inlet for injection. These procedures provide alternatives when some paths become unavailable.
• Steam-driven RCIC or HPCI is one of few available systems during the initial phase of station blackout. There are accordingly many paths generated using RCIC or HPCI pump to provide pressure (Procedures 10-17 in Table 6.3, and Procedures 4-7 in Table 6.4). It is thus crucial to maintain RCIC or HPCI available, which is also indicated by Strategy-3. The main concern is how the pump trip can be overridden by design of the control logic.
• Strategy-4andStrategy-5involve extending battery life as much as possible and how to recharge battery to maintain DC loads for control. Note that DC power is necessary for most of mitigation strategies. Loss of DC power will result in unavailability of all operable components. Since MFM primarily describes function of component, and the MFM-based planning technique expects to find operations on components that can affect functions, it is hardly to generate such strategies when no DC power is available for component operations.
• Strategy-8mentions to use CRD as an approach for vessel injection, which is also generated by Procedure-6 in Table 6.3, although CRD is not originally designed for injection. The water source for CRD is CST, whose importance has been emphasized inStrategy-1. The power source can be changed with a mobile diesel generator.
• Strategy-9emphasizes to use FP as a strategies for vessel injection and containment spray. There are pipeline connection between FP and the other systems. The water source can be infinite and the power by fire engines can be relatively easier brought from outside. Many procedures using FP are also generated by MFM-based technique (Procedures-5, 9, and 20 in Table 6.3, and Procedure-2 in Table 6.4).
S/P
CST
FWT
SLC
Normal AC power
Emergency AC power
Cross-tie
AC power Fire
engine Mobile diesel
generator Steam
Modeling boundary
Power sources Water
sources
MFM model of BWR
4 path
3 path 1 path 4 path
3 path
8 path 1 path
3 path 3 path
4 path
3 path
0 path 8 path
3 path
Goal of preventing core damage
Maintenance
Supplement
Fig. 6.5 MFM capability for planning mitigation strategies in the BWR case
• There are several SAM strategies, which are also hardly generated by MFM.Strategy-6 involves maintaining control availability of SRVs, which is also important for Strategy-10. Strategy-7involves how to apply AC power from the other units in the power station.
According to the comparison results, in the following, capabilities as well as restrictions of MFM for planning mitigation strategies are discussed. Figure 6.5, showing a summary of the generated mitigation strategies for achieving core damage prevention, will be used for demonstration. Apart from three procedures, which only rely on DC power for operation actuation, all generated procedures need water and power sources. In other words, each path implied by corresponding procedure must have connections with both asourcefunction representing water source and anothersourcefunction representing power source. Note that the current model of BWR sets thosesourcesas modeling boundary.
• It is advisable to provide operators alternative options of mitigation strategies as much as possible. As shown in Fig. 6.5, MFM has capability of planning alternatives.
• The MFM-based OPS approach devotes to identifying available equipment to establish valid paths that can achieve the goal. Loss of DC power implies that equipment will no longer be available for control. In terms of MFM, the corresponding function will lose potentiality, which can be not directly represented by the current MFM concept. In this sense, MFM has no capability of planning what absent in the model.
• What can be planned by MFM heavily depends on settings of the modeling boundary and the abstraction levels being chosen for modeling. Since the model of BWR sets water and power sources as the boundary, it is hard to generate strategies for solving problems of both two sources, as those mentioned in Table 6.5, i.e. methods of supplying water and maintaining power.
• However, the results can provide insights on how above sources should be planned in advance. For instance, there are eight paths using CST as the water sources, and it emphasizes the importance of CST, which has been implied inStrategy-1. As for the power source, generating eight procedures using steam-driven systems is self-evident that RCIC and HPCI are very important for maintaining the injection function, which is also implied inStrategy-3.
Conclusions and future works
It is desired to develop a decision-making support system to facilitate operators’ performance of knowledge-based planning. The difficulties faced by human in building mental models of plant and reasoning plans can be solved by computer’s capabilities of knowledge rep-resentation and systematic inference. By literature review, it is argued that most existing operating procedure synthesis (OPS) systems fail to explicitly and adequately describe plant knowledge. This restricts capabilities of planning responding strategies involving alternative systems or those that can achieve functions that may not be typically achieved in that way, which are commonly required in situations that call upon the knowledge-based planning.
Multilevel flow modeling (MFM), whose way of decomposing and integrating plant knowl-edge is considered to be consistent with the standard function-oriental emergency procedure development process. Besides, MFM is able to cover both intention and causality aspects of actions, which makes MFM suitable and competitive to be the domain knowledge bases for the action planning. This thesis presents an application of MFM in the field of OPS. A case in the nuclear domain is used to demonstrate capabilities and limitations of the proposed MFM-based planning support technique. In this chapter, main contributions and potential future works extended from this thesis are briefly summarized.