No description available.

No description available.

No description available.

published_or_final_version / Industrial and Manufacturing Systems Engineering / Master / Master of Philosophy

Well designed experiments can dramatically improve the statistical accuracy of simulation output data and facilitate the statistical analysis. In practice, however, the design of experiments is the most neglected phase of a simulation study. What is needed is good, sound experimental design strategies that can be easily implemented by practitioners. In this dissertation, an experimental design strategy is proposed for finding robust system designs in discrete-event simulation, where robust system designs are system designs that are insensitive to uncontrollable system noise. A conceptual framework for automating the proposed experimental design strategy is constructed and a prototype system for testing the efficacy of the proposed strategy is developed. Concentration is on the domain area of manufacturing systems. The experimental design strategy proposed in this dissertation can be broken down into three phases. The first phase involves finding a range of feasible values for the system of design variables. This is accomplished by augmenting the simulation program with an expert system. The second phase involves constructing experimental design plans. The experimental design plans suggested are plans that allow robust system designs to be found. The third phase, the optimization phase, involves searching for a best system design, where a best system design is a system design which optimizes the performance measures of interest and is not sensitive to uncontrollable system noise. The strategy proposed is evaluated with two models of a jobshop.

No description available.

Thesis--Ph. D., University of Hong Kong, 1987.

Pre-tests and post tests were used to assess the effectiveness of an engineering high school unit on experimental design and data driven design. The engineering data acquisition unit examined in this report used project based learning to teach the design of an engineering experiment and data driven design as part of the engineering design process. The project consists of the design of a building that can safely withstand an earthquake. Students construct, test and collect data on baseline buildings, with and without load using a shaker table and data acquisition. Students' then design experiments to evaluate design modifications that will meet the customer's needs. Overall, although the number of participants was limited, the survey instruments indicated that understanding of experimental design improved among high school students participating in the unit. Based on this pilot implementation of survey instruments, some of the survey questions were clarified. / text

The intention with this thesis was to evaluate how much pickling that have been achieved on aluminium details in the pickling bath before the anodization in an anodization process. As there were no earlier studies to use around pickling before anodization, statistical experimental design was chosen as planning tool. Statistical experimental design was used to plan the experiments in an organized way and to evaluate how the pickling process works together with the main effects and the interaction effects. Detailed scientific studies were performed on how aluminium is prepared and how the anodization process works. The studies were performed in reference books. Three different methods were designed. The first method, method 1, was based on a fractional factorial design with four design variables, temperature, sodium hydroxide and aluminium concentrations and the time the details was submerged into the pickling bath. The aluminium details was made from a square profile pipe. There was nine experiments performed in method 1. The measurements on the pickling was performed in two ways, first with a dial indicator where the pickling was compared with a reference surface before and after, and another method also performed with a dial indicator, where the measurements was performed over the edge between the pickled surface and the reference surface. A statistical control calculation was done on the surface smoothness of the square profile pipes. The control showed that the standard deviation was 11 µm. Method 2 was based on a complete factorial design where the design variables was temperature and the time the details was submerged into the pickling bath. All aluminium details were homogenous. There were seven experiments performed in method 2. The measurements on the pickling was performed in two ways, first with a dial indicator where the pickling was compared with a reference surface before and after, and another method also performed with a dial indicator, where the measurements was performed over the edge between the pickled surface and the reference surface. A statistical control calculation was done on the surface smoothness of the homogenous details. The control showed that the standard deviation was 14 µm. Method 3 was designed in a different way than method 1 and 2. In method 3 one experiment was performed and the design variable which was changed was the time when the details was submerged into the pickling bath. The aluminium details had the form of homogenous cubes. The measurements on the pickling was performed by measuring the weight of the details on an analytical scale before and after the pickling, and then calculate the pickling in µm in two different ways. The first way was to use the atomic radius of aluminium and the second way was to use the size of the unit cell of aluminium. The two first methods gave very different results than the third method. The result for method 1 showed very random values with great dispersion which resulted in a non-detectable pickling. The result for method 2 was very similar to the result from method 1, very random values with great dispersion and no pickling was detectable with any confidence. The result from method 3 gave a theoretical calculated result for the pickling, when the aluminium details was submerged in the pickling bath for 1 minute, and based on the atomic radius of aluminium, of 1,52 µm and with the same conditions but using the unit cell of aluminium showed a pickling of 1,62 µm. When the aluminium details were submerged in the pickling bath for 3 minutes, the theoretical calculation with the atomic radius of aluminium gave that the pickling was 4,51 µm and with the same conditions but using the unit cell of aluminium showed a pickling of 4,79 µm.

No description available.