[en] The present work shows the results obtained after applying the Bee Colony Optimization algorithm in the design of fuel cells for a BWR. The algorithm that is implemented, works following the behavior that have the bees when pollinating a flowers field. The bees carry out an exhaustive analysis in the cell, so they leave generating diverse configurations where different fuel bars are placed with different uranium enrichments to reach a value mean, with a specific number of gadolinium bars. The behavior of the generated cell is evaluated by means of the use of the commercial code CASMO-4, which shows the variables that allow fixing if the cell fulfills the requirements. Such variables are the local potential peak factor and the neutrons multiplication factor in an infinite medium. (Author)

[en] In this paper a new method for the pre-design of a typical fuel cell with a structural array of 10 x 10 fuel elements for a BWR is presented. The method is based on principles of maximum dispersion and minimum peaks of local power within the array of fuel elements. The pre-design of the fuel cells is made by simulation in two dimensions (2-D) through the cells physics code CASMO-4. For this purpose of pre-design the search process is guided by an objective function which is a combination of the main neutronic parameters of the fuel cell. The results show that the method is a promising tool that could be used for the design of fuel cells for use in a nuclear plant BWR. (Author)

[en] The objective of the report was to carry out a comparative analysis of costs of energy generation among the designs GE9B of General Electric, 9X9-IX of SIEMENS and SVEA-96 of ABB ATOM, proposed to be used as recharge fuel in the Unit 2 of the Laguna Verde Nuclear Power station. (Author)

[en] In this work the results obtained when analyzing new strategies in the reload designs of nuclear fuel and models of control bars, for boiling water reactors are presented. The idea is to analyze the behaviour of the reactor during an operation cycle, when the heuristic rules are not used (commonly used by expert engineers in both designs). Specifically was analyzed the rule of low leak and the load strategy Control Cell Core for the design of a fuel reload. In a same way was analyzed the rule of prohibiting the use of the intermediate positions in the control bars, as well as the construction of bar models based on load strategies type Control Cell Core. In the first analysis a balance and transition cycle were used. For the second analysis only a transition cycle was used, firstly with the reloads designed in the first analysis and later on with reloads built by other methods. For the simulation of the different configurations proposed in both cases, was used the code Simulate-3. To obtain the designs in both studies, the heuristic techniques or neural networks and taboo search were used. The obtained results show that it can be omitted of some rules used in the ambit for the mentioned designs and even so to obtain good results. To carry out this investigation was used Dell work station under Li nux platform. (Author)

[en] In this work are presented the obtained preliminary results to design nuclear fuel cells for boiling water reactors (BWR) using new strategies. To carry out the cells design some of the used rules in the fuel administration were discarded and other were implemented. The above-mentioned with the idea of making a comparative analysis between the used rules and those implemented here, under the hypothesis that it can be possible to design nuclear fuel cells without using all the used rules and executing the security restrictions that are imposed in these cases. To evaluate the quality of the obtained cells it was taken into account the power pick factor and the infinite multiplication factor, in the same sense, to evaluate the proposed configurations and to obtain the mentioned parameters was used the CASMO-4 code. To optimize the design it is uses the combinatorial optimization technique named Path Re linking and the Dispersed Search as local search method. The preliminary results show that it is possible to implement new strategies for the cells design of nuclear fuel following new rules. (Author)

[en] An approach to solving the problem of fuel cell design for BWR power reactor is presented. For this purpose the hybridization of a method based in heuristic knowledge rules called S15 and the advantages of a meta-heuristic method is proposed. The synergy of potentialities of both techniques allows finding solutions of more quality. The quality of each solution is obtained through a multi-objective function formed from the main cell parameters that are provided or obtained during the simulation with the CASMO-4 code. To evaluate this alternative of solution nuclear fuel cells of reference of nuclear power plant of Laguna Verde were used. The results show that in a systematic way the results improve when both methods are coupled. As a result of the hybridization process of the mentioned techniques an improvement is achieved in a range of 2% with regard to the achieved results in an independent way by the S15 method. (Author)

[en] At present the techniques of evolution al computation receive an increasing attention in the scientific and technological areas. This situation is due to its enormous potential in the optimization applied to problems of discussed computational complexity. In the nuclear area these techniques are used in diverse problems of combinatory optimization related with designing cores of power reactors. A distinctive characteristic of the evolution al and/or meta-heuristic algorithms is that appeal in each one from their applications to an adjustment function, fitness or of quality. This function allows to discriminate or to evaluate potentials solutions of the problem to solve. The definition of this situation is very important since it allows following the search of the algorithm toward different regions of the search space. In this work the impact that has the election of this function in the quality of the found solution is shown. The optimization technique by ant colonies or Acs (ant colony system) was used applied to the radial design of fuel cells for a boiling water power reactor. The notable results of the Acs allowed to propose the adjustment method of the importance and with this to obtain adjustment functions that guide the search of solutions of collective algorithms efficiently, basic capacity to develop the proposal of emulation of the natural selection and to investigate the possibility that on order of specify goals, to obtain the corresponding decision variables. A variety of re tro-exit (re tro-out) complementary process of feedback (re tro-in) that opens extended application perspectives of be feasible. (Author)

[en] In this paper the simulation of a BWR in order to evaluate the performance of a set of fuel assemblies under stationary state in three dimensions (3-D) is presented. 15 cases selected from a database containing a total of 18225 cases are evaluated. The main selection criteria were based on the results of the design phase of the power cells in two dimensions (2-D) and 3-D initial study. In 2-D studies the parameters that were used to qualify and select the designs were basically the local power peaking factor and neutron multiplication factor of each fuel cell. In the initial 3-D study variables that defined the quality of results, and from which the selection was realized, are the margins to thermal limits of reactor operation and the value of the effective multiplication factor at the end of cycle operation. From the 2-D and 3-D results of the studies described a second 3-D study was realized, where the optimizations of the fuel reload pattern was carried out. The results presented in this paper correspond to this second 3-D study. It was found that the designs of the fuel cell they had a similar behavior to those provided by the fuel supplier of reference BWR. Particularly it noted the impact of reload pattern on the cold shut down margin. An estimate of the operation costs of reference cycle analyzed with each one designed reload batch was also performed. As a result a positive difference (gain) up to 10,347 M/US D was found. (Author)

[en] Preliminary results are presented for the design of nuclear fuel cells, used in the design of new fuel assemblies. To obtain these results, heuristic techniques known as neural networks and a hybrid between scattered search and trajectory re-chaining were implemented. Because the computational requirements are large, it was necessary to implement parallel computing tools to reduce response times. The objective function implemented includes the local peak power factor of the cell throughout its life, as well as the infinite neutron multiplication factor. The enrichment of ²³⁵U pills increased, beyond the usual, allowing pills with up to 5.9% enrichment. The designed cells were used to propose new fuel assemblies for BWR reactors. To test our designs, the fuel assemblies were used in an optimization process to generate the refills and the respective control bar patterns. Our results show that it is possible to increase the bank of pills available for the design of cells and design refills that comply with the security restrictions and the requested energy requirements. (Author)

[en] This work presents the preliminary results of a statistical analysis carried out for the design of nuclear fuel cells. The analysis consists in verifying the behavior of a cell, related with the frequency of the pines used for its design. In this preliminary study was analyzed the behavior of infinite multiplication factor and the peak factor of local power. On the other hand, the mentioned analysis was carried out using a pines group of enriched uranium previously established, for which varies the pines frequency used in the design. To carry out the study, the CASMO-IV code was used. The obtained designs are for the different axial areas of a fuel assembly. A balance cycle of the unit 1 of the nuclear power plant of Laguna Verde was used like reference. To obtain the result of the present work, systems that are already had and in which have already been implemented the heuristic techniques of ant colonies, neural networks and a hybrid between the dispersed search and the trajectories re-chaining. The results show that is possible to design nuclear fuel cells with a good performance, if is considered a statistical behavior in the frequency of the used pines, in a same way. (Author)