No description available.

No description available.

The purpose of this thesis is to improve the 'art' of early capital cost estimation of chemical process plants. Capital cost estimates are required in the early business planning and feasibility assessment stages of a project, in order to evaluate viability and to compare the economics of the alternative processes and operating conditions that are under consideration for the plant. There is limited knowledge about a new plant in the early stages of process development. Nevertheless, accurate cost estimates are needed to prevent incorrect decisions being made, such as terminating the development of a would-be profitable plant. The published early capital cost estimation methods are described. The methods are grouped into three types of estimate: exponent, factorial and functional unit. The performance of these methods when used to estimate the capital costs of chemical plants is assessed. A new estimating method is presented. This method was developed using the same standard regression techniques as used in the published methods, but derived from a new set of chemical plant data. The effect that computers have had on capital cost estimating and the future possibilities for the use of the latest computer techniques are assessed. This leads to the fuzzy matching technique being chosen to develop a new method for capital cost estimation. The results achieved when using fuzzy matching to estimate the capital cost of chemical plants are presented. These results show that the new method is better than those that already exist. Finally, there is a brief discussion of how fuzzy matching could be applied in the future to other fields of chemical engineering.

Selective laser sintering (SLS) is a rapid prototyping process, which operates by using a laser to locally heat an area within a layer of powder material, causing it to fuse together, creating a thin cross-section of solid material. 3D shapes are built by repeatedly depositing a layer of fresh powder on top of the cross-section and then locally heating it, causing it to fuse together and to the layer beneath. However, during SLS processing, temperature differences that exist in different regions of the fabricated parts lead to uneven shrinkages. The shrinkages cause surfaces in the part, which are intended to be flat, to exhibit a curved profile; a phenomenon termed curl. The development of curl is highly influenced by the SLS machine parameters selected in fabrication. The production of geometrically acceptable parts involves numerous fabrication trials before the optimum machine parameters can be found. The procedure can be time consuming and expensive. The aim of the work presented in this thesis was to develop finite element models for the purpose of predicting curl in SLS fabricated polycarbonate parts. The ultimate goal was to use the models to estimate the optimum SLS machine parameters for the physical fabrication of geometrically acceptable parts, produced in any material, and therefore avoid the costly and time consuming process of using SLS machines for experimental purposes. The prediction of curl was made through heat transfer and stress finite element models that were both coupled using the sequentially coupled thermal-stress analysis technique. Experimental work was carried out to measure material properties used as input to the models and to validate results predicted. The sensitivity of curl predicted to assumptions considered in the heat and stress models was introduced, and the assumptions highly influencing the accuracy of curl predictions were identified.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.