Earl E. Feldman

The experimental breeder reactor II inherent shutdown and heat removal tests — results and analysis

Abstract A test program is being conducted to demonstrate that a power-producing liquid-metal reactor (LMR) can (1) passively remove shutdown heat by natural convection, (2) passively reduce power in response to a loss of reactor flow, and (3) passively reduce power in response to a loss of the balance-of-plant heat sink. Measurements and pretest predictions confirm that natural convection is a reliable, predictable method of shutdown heat removal and suggest that safety-related pumps or pony motors are not necessary for safe shutdown heat removal in an LMR. Measurements from tests in which reactor flow and heat rejection to the balance of plant were perturbed show that reactivity feedbacks can passively control power and temperature. Data from these tests form a basis for additional tests including a complete loss of flow without scram and a complete loss of heat sink without scram.

Loss-of-primary-flow-without-scram tests: Pretest predictions and preliminary results

Abstract A series of tests in the Experimental Breeder Reactor No. 2 (EBR-II) has been concluded that investigated the effects of a complete loss of primary flow without scram. The development and preliminary study of these events is first discussed, including the test limits and controlling parameters. The results of two of the tests, SHRT 39 and 45, are examined in detail, although a compact summary of all the tests is included. The success in meeting the objectives of the test program served to verify that natural processes will shut down the reactor and maintain adequate cooling without control rod or operator intervention. The good comparison between predicted and measured results confirms that such events can be analyzed without elaborate codes if the basic processes are understood. Furthermore, recent studies suggest that the EBR-II results are characteristic of new innovative LMR designs being pursued in the U.S. that incorporate metallic driver fuel.

Identification and control of NOx emissions using neural networks

We investigate the application of two classes of artificial neural networks for the identification and control of discrete-time nonlinear dynamical systems. A fully connected recurrent network is used for process identification, and a multilayer feedforward network is used for process control. The two neural networks are arranged in series for closed-loop control of oxides of nitrogen (NOx) emissions of a simplified representation of a dynamical system. Plant data from one of Commonwealth Edisons coal-fired power plants are used for testing the approach, with initial results indicating that the method is feasible. However, further work is required to determine whether the method remains feasible as the number of state variables and control variables are increased.

Multilayer perceptron for nonlinear programming

A new method for solving nonlinear programming problems within the framework of a multilayer neural network perceptron is proposed. The method employs the Penalty Function method to transform a constrained optimization problem into a sequence of unconstrained optimization problems and then solves the sequence of unconstrained optimizations of the transformed problem by training a series of multilayer percepperptrons. The neural network formulation is represented in such a way that the multilayer perceptron prediction error to be minimized mimics the objective function of the unconstrained problem, and therefore, the minimization of the objective function for each unconstrained optimization is attained by training a single perceptron. The multilayer perceptron allows for the transformation of problems with two-sided bounding constraints on the decision variables x, e.g., a ≤ xn ≤ b, into equivalent optimization problems in which these constraints do not explicitly appear. Hence, when these are the only constraints in the problem, the transformed problem is constraint free (i.e., the transformed objective function contains no penalty terms) and is solved by training a multilayer perceptron only once. In addition, we present a new Penalty Function method for solving nonlinear programming problems that is parameter free and guarantees that feasible solutions are obtained when the optimal solution is on the boundary of the feasible region. Simulation results, including an example from operations research, illustrate the proposed methods.

EBR-II unprotected loss-of-heat-sink predictions and preliminary test results☆

Abstract Two unprotected (i.e., no scram or plant protection system action) loss-of-heat-sink transients were performed on the Experimental Breeder Reactor-II in the Spring of 1986. One was initiated from full power (60 MW) and the other from half power. The loss of heat sink was accomplished in each test by essentially stopping the secondary-loop sodium coolant flow. Pretest predictions along with preliminary test results demonstrate that the reactor shuts itself down in a benign and predictable manner in which all of the reactor temperatures approach a quenching (or smothering) temperature at which the fission power goes to zero.

Safety analysis for the loss-of-flow and loss-of-heat sink without scram tests in EBR-II☆

Abstract This paper discusses the safety considerations and the analysis that was done to support the conduct of the Loss-of-Flow and Loss-of-Heat Sink Without Scram tests in EBR-II. The plant safety limits and the analysis codes and methods are presented. The impact of the tests and off-normal conditions on the fuel and plant structures is evaluated. Conclusions are that the test program had no impact on plant operations and the accumulated fuel damage was negligible.

Demonstration of EBR-II power maneuvers without control rod movement

Abstract A group of five plant inherent control tests was successfully conducted in November 1987 in the Experimental Breeder Reactor II. These tests demonstrated that the plant power of a metal-fueled reactor can be passively controlled over a large power range by slowly changing the primary flow and the reactor inlet temperature. These variables are, in turn, regulated by the primary pump speed, the secondary flow, and the turbine inlet pressure. In all tests, control rods were not used to regulate power. It was demonstrated that the plant power can be controlled with reasonable accuracy without using control rods when the reactivity feedback characteristics of the reactor are well understood and the plant controllers are adequately designed.

Nonlinear programming with feedforward neural networks

We provide a practical and effective method for solving constrained optimization problems by successively training a multilayer feedforward neural network in a coupled neural-network/objective-function representation. Nonlinear programming problems are easily mapped into this representation which has a simpler and more transparent method of solution than the optimization performed with Hopfield-like networks and poses very mild requirements on the functions appearing in the problem. Simulation results are illustrated and compared with an off-the-shelf-optimization tool.

An intelligent emissions controller for fuel lean gas reburn in coal-fired power plants.

ABSTRACT The application of artificial intelligence techniques for performance optimization of the fuel lean gas reburn (FLGR) system is investigated. A multilayer, feedforward artificial neural network is applied to model static nonlinear relationships between the distribution of injected natural gas into the upper region of the furnace of a coal-fired boiler and the corresponding oxides of nitrogen (NOx) emissions exiting the furnace. Based on this model, optimal distributions of injected gas are determined such that the largest NOx reduction is achieved for each value of total injected gas. This optimization is accomplished through the development of a new optimization method based on neural networks. This new optimal control algorithm, which can be used as an alternative generic tool for solving multidimensional nonlinear constrained optimization problems, is described and its results are successfully validated against an off-the-shelf tool for solving mathematical programming problems. Encouraging results obtained using plant data from one of Commonwealth Edisons coal-fired electric power plants demonstrate the feasibility of the overall approach. Preliminary results show that the use of this intelligent controller will also enable the determination of the most cost-effective operating conditions of the FLGR system by considering, along with the optimal distribution of the injected gas, the cost differential between natural gas and coal and the open-market price of NOx emission credits. Further study, however, is necessary, including the construction of a more comprehensive database, needed to develop high-fidelity process models and to add carbon monoxide (CO) emissions to the model of the gas reburn system.

Results and Implications of the Experimental Breeder Reactor II Inherent Safety Demonstration Tests

Two milestone tests were conducted in the Experimental Breeder Reactor II (EBR-II), demonstrating some of the inherent safety features of a liquid-metal reactor. The first test was a loss of flow without scram and the second was a loss of heat sink without scram. Both tests were initiated from 100% power, and in both tests the reactor was shut down passively - by natural processes, principally thermal expansion - without automatic scram, operator intervention, or the help of special incore devices. The temperature transients during the tests were mild, as predicted, and there was no damage to the core or reactor plant structures. The test plus analysis demonstrated the feasibility of inherent passive shutdown for under cooling accidents and identified the more important features necessary for inherent shutdown and passive cooling. The results provide a technical basis for future experiments in EBR-II to investigate inherent safety for transient overpower accidents and to provide additional data for validation of computer codes used for design and safety analysis of inherently safe reactor plants.