Latent heat during recalesence and nucleation and post-recalesence temperature was analysed for refractory metals. An effect on latent heat was found by alloying the pure elements with other refractories. Latent heat was found to be 15-65% of tabulated values. Interface energy was evaluated from undercooling experiments. It was found that the dimensionless numbers β (σLs/ σLv) and α can be used to classify elements into distinctive groups and crystallographic structure. The phase diagrams for Al-Ti, Al-Ti-B and Al-Ti-C as well as inoculation were analysed. It was found that Al nucleates on Al3Ti which nucleates on TiB2. TiC was found to decompose into Al4C3 and Al3Ti. The inoculation of nodular cast iron with Mg, Ce; Ca and the formation of sulphides and oxides was analysed. The formation of new inclusions during the solidification as well as the formation of graphite is discussed. / QC 20100602

Wireless Sensor Networks (WSNets), consisting of a lot of Wireless Sensor Nodes (WSNs), play an important role in structural health and machine condition monitoring. But the WSNs provided by the current market cannot meet the diversity of application requirements because they have limited functions, unreliable node performance, high node cost, high system redundancy, and short node lifespan. The aim of the research is to design the architecture of a WSN with low power consumption and node cost, which can be dynamically configured according to application requirements for structural health and machine condition monitoring. This research investigates the improvement of node performance and reliability through the new design methodologies and the extension of node lifespan by interfacing energy harvesters and implementing node power management. The main contributions of the research are presented from the following aspects:1. Model development of node architecture for application diversityThe merits of model include: (1) The proposed node architecture can be dynamically configured in terms of application requirements for reducing system redundancy, power consumption and cost; (2) It supports multichannel signal measurement with the synchronous and asynchronous signal sampling modules and three interface circuits; (3)The model parameters can be calculated; (4) As the model is based on discrete electronic components, it can be implemented by using Components-Off-The-Shelf (COTS).2. A novel pipeline design of the built-in ADC inside a microprocessorThe merit of proposed pipeline solution lies in that the sampling time of the built-in ADCs is reduced to one third of the original value, when the ADC operates in sequence sampling mode based on multichannel signal measurement.3. Self-adjusting measurement of sampled signal amplitude This work provides a novel method to avoid the distortion of sampled signals even though the environmental signal changes randomly and over the sampling range of the node ADC. The proposed method can be implemented with four different solutions.4. Interface design to support energy harvesting The proposed interface will allow to: (1) collect the paroxysmal ambient energy as more as possible; (2) store energy to a distribution super-capacitor array; (3) harvest electrical energy at high voltage using piezoelectric materials without any transformer; (4) support the diversity of energy transducers; and (5) perform with high conversion efficiency.5. A new network task scheduling model for node wireless transceiver The model allows to: (1) calculate node power consumption according to network task scheduling; (2) obtain the optimal policy for scheduling network task.6. A new work-flow model for a WSN The model provides an easy way to (1) calculate node power consumption according to the work flow inside a WSN; (2) take fully advantage of the power modes of node electronic components rather than outside factors; (3) improve effectively node design.

The temperature distribution of the lubricating interfaces is an important aspect of the functioning of positive displacement machines. It can determine the efficiency and the life time of the unit. In particular, it directly affects the fluid properties and the thermal induced deformations of the solid bodies. A simulation tool capable of predicting the fluid temperature in such gaps thus becomes very useful in the design process of these machines. The temperature distribution in a film comprises of many physical phenomena including convection and conduction along and across the film. Past numerical approaches solved this multi-directional conduction-convection problem using a threedimensional(3D) grid, making the tool computationally expensive and unsuitable for fast simulations. This paper proposes a hybrid fluid temperature solver, based on, a low computational cost twodimensional(2D) grid, to reduce the simulation time with reasonable accuracy. The piston/cylinder interface of an axial piston machine is selected as reference case to demonstrate the proposed approach. The hybrid approach was found to speedup the simulation times by 36%.