• No results found

Chapter 1 Introduction

1.3 Objectives and Outline

source. Nuclear aerosols tend to be charged due to transporting radioactive isotopes that emit  and  radiation carrying away electric charge.

1.2.2.3 Particle Resuspension

In reactors, particles deposited on surfaces can become resuspended by an aerodynamically- generated lift and drag forces due to the incidents already described above. The basic nature of the resuspension process is understood but it is not possible with today’s models to predict reliably when and to what extent the process occurs (see Chapter 2).

Many parameters need to be considered in the resuspension process such as particle size, flow conditions and the adhesion strength between particles and wall. Surface roughness is a significant parameter in adhesive force. Resuspension is a random process. Particles are not immediately resuspended by a certain flow but by irregular bursts. In turbulent flow, small particles are resuspended in the boundary layer due to the occasional bursts of turbulent eddies. Resuspension is a key precursor to dust explosions (cf. ITER above). In nuclear accidents, resuspension can have considerable impact on the progression of the accident and radiation releases (Williams & Loyalka, 1991).

1.3 Objectives and Outline

1.3.1 Objectives

The underlying objective of this research is to develop, validate and apply a physical model for the resuspension of particles from multilayer deposits that can account for the resuspension of clusters of particles from a bed of particles of variable size and shape. It was not expected that a final model taking account of an arbitrary deposit in arbitrary conditions would be developed but that very substantial progress would be made towards this objective. The practical application will be that of assessing resuspension of particles in a range of accident scenarios for PWRs as well as other reactor technologies such as ITER. It will be seen that progress has been made with a hybrid development of the Rock’n’Roll (Reeks & Hall, 2001) model suited to application to multilayer deposits.

The resuspension of a particle from a multilayer deposit depends on the interaction of two processes: the aerodynamic forces tending to remove the particle from the bed and the various forces (essentially due to particle-particle interactions) that resist resuspension. To some extent, these two effects can be decoupled and studied separately. The aerodynamic forces acting on a particle will depend mainly on the turbulent flow and will be largely independent of the nature of the surface so that results obtained for an isolated particle on a smooth wall can be applied to an

1.3 Objectives and Outline

exposed particle sitting at the surface of a multilayer deposit. The form of the particle resistance to resuspension that is assumed in the Rock’n’Roll model might also be expected to apply to exposed particles on a multilayer deposit. But there are also good physical reasons for believing that other effects might occur in the resuspension of particles from a multilayer model that are not present in the resuspension of an isolated particle from a smooth wall. For example, the micro- and macro- scale roughness of the multilayer deposit will have some influence on the turbulent structure of the near-wall boundary layer, thus modifying the aerodynamic forces acting on the particle. And the presence of a wide range of particle sizes within the multilayer deposit can result in additional resistance forces caused by interaction between the particles. An important consequence of this interaction is that the resistance to resuspension can vary in time – the largest particles will tend to disappear first, leaving a high concentration of small particles in the surface layer. These small particles might then act as ‘cement’, locking in the larger particles in the lower layers, and thus modifying the threshold for particle erosion. In such a situation the Rock’n’Roll model might no longer be a reliable or relevant model for the intervening physical processes.

There remain some physical uncertainties in the basic Rock’n’Roll model, notably concerning the validity of the assumptions made in modelling the aerodynamic forces on a particle at the surface. It was decided, therefore, that since the Rock’n’Roll model will form the basis of a model for resuspension from a multilayer deposit, the first priority of this research would be to improve the modelling of the turbulence-induced aerodynamic forces on an isolated particle. It was thought that investigating how the presence of layers of particles might modify the turbulence and the aerodynamic forces exerted on the particle might also be achieved. Some account of how the resistance to resuspension is modified for particles at the surface of a multilayer deposit was also a possibility. The results from these possibilities could then be used to develop a simpler engineering model for resuspension from multilayer deposits.

Detailed objectives begin with improving the Rock’n’Roll model for the resuspension of particles from surfaces with less than a mono-layer coverage. Simulations will be required of the stochastic properties of the turbulent forces generated on particles due to the turbulent structures generated in the boundary layer (sweeping and bursting mechanisms). In particular, the examination of the sensitivity or otherwise of the Gaussian assumption used in earlier stochastic models for resuspension must be verified. This data will be used to construct a model for resuspension rate that takes account of the non-Gaussian nature of the turbulent fluctuations (in particular the turbulent induced shear force). It will be important to see how, for instance, the spectrum of the turbulent-fluctuation strain rates is coupled to the induced aerodynamic forces. These forces will be calculated using various models for the lift and drag based on the local instantaneous strain rate and fluid velocity. Then, the intention was to develop a mechanistic model for resuspension of particles of variable sizes (and perhaps shapes) from multilayer deposits accounting for the resuspension of particles in clusters. A detailed model must deal with detailed mechanisms for

1.3 Objectives and Outline

release of clusters under the actions of turbulent drag and lift forces and their couples and the cohesive forces binding the deposit together.

1.3.2 Outline

There are six chapters in this thesis. The present Chapter is the introduction which describes the background of severe nuclear accidents and aerosol issues in different types of reactors as well as the objective of this research. Chapter 2 is a literature review which presents the previous theories on forces acting on particles when they are near the boundary layer, classical resuspension models and previous resuspension experiments. The previous classical resuspension models are summarized in three parts: 1) models based on force balance, 2) models based on energy accumulation, 3) multilayer-deposit resuspension. Modifications of resuspension models rendering them physically more acceptable will be demonstrated in Chapter 3 and Chapter 4, i.e., Chapter 3: modification of a typical force-balance model (the fluctuating aerodynamic forces will be considered); Chapter 4: modification of an energy-accumulation model (the statistics of fluid fluctuations will be modified based on Large Eddy Simulation and Direct Numerical Simulation data). In Chapter 5, a multilayer resuspension model based on energy accumulation will be developed, tested and discussed. Finally, Chapter 6 leads to the final conclusions and future possibilities.

Related documents