Abstract
A high-fidelity three-dimensional space time nodal method has been developed to simulate the dynamics of the reactor core for real time simulation. This three-dimensional reactor core mathematical model can be composed of six sub-models, neutron kinetics model, cay heat model, fuel conduction model, thermal hydraulics model, lower plenum model, and core flow distribution model. During simulation of each sub-model some operation data will be produced and lots of valuable, important information reflecting the reactor core operation status could be hidden in, so how to discovery these information becomes the primary mission people concern. Under this background, data mining (DM) is just created and developed to solve this problem, no matter what engineering aspects or business fields. Generally speaking, data mining is a process of finding some useful and interested information from huge data pool. Support Vector Machine (SVM) is a new technique of data mining appeared in recent years, and SVR is a transformed method of SVM which is applied in regression cases. This paper presents only two significant sub-models of three-dimensional reactor core mathematical model, the nodal space time neutron kinetics model and the thermal hydraulics model, based on which the neutron flux and enthalpy distributions of the core are obtained by solving the three-dimensional nodal space time kinetics equations and energy equations for both single and two-phase flows respectively.Moreover, it describes that the three-dimensional reactor core model can also be used to calculate and determine the reactivity effects of the moderator temperature, boron concentration, fuel temperature, coolant void, xenon worth, samarium worth, control element positions (CEAs) and core burnup status. Besides these, the main mathematic theory of SVR is introduced briefly next, on the basis of which SVR is applied to dealing with the data generated by two sample calculation, rod ejection transient and axial flux offset. Comparing with the traditional data analysis method, the main character of SVR is not enforcing a priori information or judgment before analysis begins, but obtaining the fitting result from inner relationship of data completely, which represents the main principle of data mining, letting the data speak themselves, and the result is more accurate and believable than traditional methods.
