No description available.

The overall goal of this doctoral dissertation is to develop a prototype instrument, a Portable Aerosol Collector and Spectrometer (PACS), that can continuously measure aerosol size distributions by number, surface area and mass concentrations over a wide size range (from 10 nm to 10 µm) while also collecting particles with impactor and diffusion stages for post-sampling chemical analyses. To achieve the goal, in the first study, we designed, built and tested the PACS hardware. The PACS consists of a six-stage particle size selector, a valve system, a water condensation particle counter to measure number concentrations and a photometer to measure mass concentrations. The valve system diverts airflow to pass sequentially through upstream stages of the selector to the detectors. The stages of the selector include three impactor and two diffusion stages, which resolve particles by size and collect particles for chemical analysis. Particle penetration by size was measured through each stage to determine actual performance and account for particle losses. The measured d50 of each stage (aerodynamic diameter for impactor stages and geometric diameter for diffusion stages) was similar to the design. The pressure drop of each stage was sufficiently low to permit its operation with portable air pumps. In the second study, we developed a multi-modal log-normal (MMLN) fitting algorithm to leverage the multi-metric, low-resolution data from one sequence of PACS measurements to estimate aerosol size distributions of number, surface area, and mass concentration in near-real-time. The algorithm uses a grid-search process and a constrained linear least-square (CLLS) solver to find a tri-mode (ultrafine, fine, and coarse), log-normal distribution that best fits the input data. We refined the algorithm to obtain accurate and precise size distributions for four aerosols typical of diverse environments: clean background, urban and freeway, coal power plant, and marine surface. Sensitivity studies were conducted to explore the influence of unknown particle density and shape factor on algorithm output. An adaptive process that refined the ranges and step sizes of the grid-search reduced the computation time to fit a single size distribution in near-real-time. Assuming standard density spheres, the aerosol size distributions fit well with the normalized mean bias (NMB) of -4.9% to 3.5%, normalized mean error (NME) of 3.3% to 27.6%, and R2 values of 0.90 to 1.00. The fitted number and mass concentration biases were within ± 10% regardless of uncertainties in density and shape. With this algorithm, the PACS is able to estimate aerosol size distributions by number, surface area, and mass concentrations from 10 nm to 10 µm in near-real-time. In the third study, we developed a new algorithm–the mass distribution by composition and size (MDCS) algorithm–to estimate the mass size distribution of various particle compositions. Then we compared the PACS for measuring multi-mode aerosols to three reference instruments, including a scanning mobility particle sizer (SMPS), an aerodynamic particle sizer (APS) and a nano micro-orifice uniform deposit impactor (nanoMOUDI). We used inductively coupled plasma mass spectrometry to measure the mass of collected particles on PACS and nanoMOUDI stages by element. For the three-mode aerosol, the aerosol size distributions in three metrics measured with the PACS agreed well with those measured with the SMPS/APS: number concentration, bias = 9.4% and R2 = 0.96; surface area, bias = 17.8%, R2 = 0.77; mass, bias = -2.2%, R2 = 0.94. Agreement was considerably poorer for the two-mode aerosol, especially for surface area and mass concentrations. Comparing to the nanoMOUDI, for the three-mode aerosol, the PACS estimated the mass median diameters (MMDs) of the coarse mode well, but overestimated the MMDs for ultrafine and fine modes. The PACS overestimated the mass concentrations of ultrafine and fine mode, but underestimated the coarse mode. This work provides insight into a novel way to simultaneously assess airborne aerosol size, composition, and concentration by number, surface area and mass using cost-effective handheld technologies.

El texto completo de este trabajo no está disponible en el Repositorio Académico UPC por restricciones de la casa editorial donde ha sido publicado. / This work proposes a low-cost hybrid hardware and software system that, through a set of methods and nested while loop fitting algorithms, allows to automatically identify the electrical and mechanical parameters of a brushed direct current motor. The aim is to facilitate a tool that contributes to the development of motion control projects in which this type of actuator is used, automating and speeding up the identification process of the motor system aiming to reach 98% accuracy, in order to guarantee a good electrical and mechanical parameter estimates for the brushed direct current motor. To achieve the objective, a platform was developed consisting of a main interface programmed in Matlab and a data acquisition hardware based on a single-phase incremental optical encoder, an H-bridge, an optocoupler circuit, and a C language-programmed DSPIC30F2010. Both parts of the platform are interconnected through the authors' own serial communication protocol.

The goal of the first project was to evaluate strategies for determining the in vitro intrinsic clearance (CLint) of dextrorphan (DR) as metabolized by the UGT2B7 enzyme to obtain dextrorphan glucuronide (DR-G). A direct injection liquid chromatography-mass spectrometry (LC-MS) method was used to monitor products using the pseudo-first-order (PFO) model. Standard enzymatic incubations were also quantified using LC-MS. These data were fit utilizing both PFO and Michaelis-Menten (MM) models to determine estimates of kinetic parameters. The CLint was determined to be 0.28 (± 0.08) µL/min/mg protein for a baculovirus insect cell-expressed UGT2B7 enzyme. This is the first confirmation that dextrorphan is specifically metabolized by UGT2B7 and the first report of these kinetic parameters. Simulated chromatographic data were used to determine the precision and accuracy in the estimation of peak volumes in comprehensive two-dimensional liquid chromatography (2D-LC). Volumes were determined both by summing the areas in the second dimension chromatograms via the moments method and by fitting the second dimension areas to a Gaussian peak. When only two second dimension signals are substantially above baseline, the accuracy and precision are poor because the solution to the Gaussian fitting algorithm is indeterminate. The fit of a Gaussian peak to the areas of the second dimension peaks is better at predicting the peak volume when there are at least three second dimension injections above the limit of detection. Based on simulations where the sampling interval and sampling phase were varied, we conclude for well-resolved peaks that the optimum precision in peak volumes in 2D separations will be obtained when the sampling ratio is approximately two. This provides an RSD of approximately 2 % for the signal-to-noise (S/N) used in this work. The precision of peak volume estimation for experimental data was also assessed, and RSD values were in the 4-5 % range. We conclude that the poorer precision found in the 2D-LC experimental data as compared to 1D-LC is due to a combination of factors, including variations in the first dimension peak shape related to undersampling and loss in S/N due to the injection of multiple smaller peaks onto the second dimension column.

Neste trabalho, apresentamos os modelos simétricos parcialmente lineares AR(1), que generalizam os modelos parcialmente lineares para a presença de erros autocorrelacionados seguindo uma estrutura de autocorrelação AR(1) e erros seguindo uma distribuição simétrica ao invés da distribuição normal. Dentre as distribuições simétricas, podemos considerar distribuições com caudas mais pesadas do que a normal, controlando a curtose e ponderando as observações aberrantes no processo de estimação. A estimação dos parâmetros do modelo é realizada por meio do critério de verossimilhança penalizada, que utiliza as funções escore e a matriz de informação de Fisher, sendo todas essas quantidades derivadas neste trabalho. O número efetivo de graus de liberdade e resultados assintóticos também são apresentados, assim como procedimentos de diagnóstico, destacando-se a obtenção da curvatura normal de influência local sob diferentes esquemas de perturbação e análise de resíduos. Uma aplicação com dados reais é apresentada como ilustração. / In this master dissertation, we present the symmetric partially linear models with AR(1) errors that generalize the normal partially linear models to contain autocorrelated errors AR(1) following a symmetric distribution instead of the normal distribution. Among the symmetric distributions, we can consider heavier tails than the normal ones, controlling the kurtosis and down-weighting outlying observations in the estimation process. The parameter estimation is made through the penalized likelihood by using score functions and the expected Fisher information. We derive these functions in this work. The effective degrees of freedom and asymptotic results are also presented as well as the residual analysis, highlighting the normal curvature of local influence under different perturbation schemes. An application with real data is given for illustration.

Neste trabalho, apresentamos os modelos simétricos parcialmente lineares AR(1), que generalizam os modelos parcialmente lineares para a presença de erros autocorrelacionados seguindo uma estrutura de autocorrelação AR(1) e erros seguindo uma distribuição simétrica ao invés da distribuição normal. Dentre as distribuições simétricas, podemos considerar distribuições com caudas mais pesadas do que a normal, controlando a curtose e ponderando as observações aberrantes no processo de estimação. A estimação dos parâmetros do modelo é realizada por meio do critério de verossimilhança penalizada, que utiliza as funções escore e a matriz de informação de Fisher, sendo todas essas quantidades derivadas neste trabalho. O número efetivo de graus de liberdade e resultados assintóticos também são apresentados, assim como procedimentos de diagnóstico, destacando-se a obtenção da curvatura normal de influência local sob diferentes esquemas de perturbação e análise de resíduos. Uma aplicação com dados reais é apresentada como ilustração. / In this master dissertation, we present the symmetric partially linear models with AR(1) errors that generalize the normal partially linear models to contain autocorrelated errors AR(1) following a symmetric distribution instead of the normal distribution. Among the symmetric distributions, we can consider heavier tails than the normal ones, controlling the kurtosis and down-weighting outlying observations in the estimation process. The parameter estimation is made through the penalized likelihood by using score functions and the expected Fisher information. We derive these functions in this work. The effective degrees of freedom and asymptotic results are also presented as well as the residual analysis, highlighting the normal curvature of local influence under different perturbation schemes. An application with real data is given for illustration.