An apparatus for cooling molten material resulting from a nuclear reactor core meltdown is disclosed. The apparatus includes a basin positioned under the reactor which is protected against excessive heat by a star-like array of heat pipes whose evaporator sections are disposed above the pan and whose condenser sections are disposed in a heat sink exterior to the containment building of the reactor. Additionally, the vertical walls of the reactor vessel chamber are similarly protected by an array of heat pipes similarly arranged and provided to intercept the radient energy of the molten core material.
A fitness function is needed for a Genetic Algorithm (GA) to work, and it appears natural that the combination of objectives and constraints into a single scalar function using arithmetic operations is appropriate. One problem with this approach, however, is that accurate scalar information must be provided on the range of objectives and constraints, to avoid one of them from dominating the other. One possible solution, then, is to try to join the objectives with the constraints with internal parameters, i.e., information that belongs to the problem itself, thereby avoiding external tuning. The building of the fitness function is so complex that, using internal or external parameters, any optimal point obtained will be a function of the coefficients used to combine objectives and constraints. However, it is possible that using internal parameters will increase performance compare to external ones.
Electrolytic cells demonstrate improved performance through use of an internal brine distribution system. A brine distributor located in the interior of the cell and positioned either at the cell bottom or above the cell's electrodes has individual brine outlets for feeding electrolyte directly to each of the cell's anolyte compartments for electrolysis. By comparison with conventional brine feed systems, e.g. cell top feed, the internal brine distributor produces higher purity gas, e.g. chlorine, at reduced power consumption and higher current efficiencies.
The Gaussian process prior formulation introduced by us in this paper learns a mapping for ordinal regression task using dual sets of latent functions. In this formulation one set of latent functions are associated with data items and the other set of latent functions are associated with entities. An entity is a term introduced by us in this work to refer to the object responsible for assigning ordinal labels to data items. For example in the collaborative filtering problem an entity corresponds to a user. In our work we assume that the entities cluster, and we use latent functions to having a Gaussian process prior to model these clusters. Similarly we also assume that the data items cluster and use latent functions having a Gaussian process prior to model these clusters. We learn the parameters of these Gaussian processes in a discriminative learning framework by minimizing a loss function using an alternating minimization procedure. The purpose of introducing dual sets of latent functions is to overcome the deficiency in the predictive nature of discriminative models for ordinal regression tasks while learning from less training data unlike generative models which have a good performance even while learning from less training data. Thus we evaluate the performance of our model on two problems, collaborative filtering and image annotation, by comparing with well known baseline methods using a generative model approach so as to understand the efficacy of our model on less training data.
