Zugl.: Berlin, Techn. Univ., Diss., 2009

Master-Arbeit Univ. St. Gallen, 2007.

Grain nitrogen (N) concentration is a major quality criterion of malting barley for which there is a narrow range that producers must meet to satisfy market requirements (1.52 – 1.84 %). In recent years growers in Ireland have had difficulty producing grain with a high enough N concentration to meet these requirements using standard recommended agronomic regimes. The reasons for the lower than expected grain N concentrations are not known. There is interest from growers and maltsters in the development of a system to forecast likely grain N concentration from crop measurements made at or before flowering. A forecasting system would allow growers to identify crops at risk of falling below specification and to apply late N fertiliser if needed. It would also enable maltsters to plan grain intake and malting operations in advance of harvest. The aim of this project was to investigate the potential for predicting grain N concentrations of spring barley from crop and soil measurements made at ear emergence. The main objectives were to 1) investigate the relationships between measurements made at ear emergence and grain N concentration at harvest in order to identify which characteristics should be included as variables in multiple regression models to explain variation in grain N concentration, 2) to use the models with independent data sets to predict grain N concentration and test the accuracy of the predictions, 3) to quantify the recovery by the crop of fertiliser N applied at anthesis and its effects on grain N concentration and 4) to determine whether non-destructive techniques can provide estimates of crop growth and N content for use in the prediction models. Field experiments were established with plots of spring barley (Hordeum vulgare cv. SY Taberna) at one site in 2013 and two sites in 2014 representative of those employed in malting barley production in Ireland. Fertiliser N applications were varied over the range 0 – 210 kg N/ha (with dressings split between sowing and mid-tillering) to provide a range of crop growth and grain N concentrations. In some experiments additional applications of N were made at anthesis to quantify effects on grain N concentration and seed rate treatments (150, 300 and 600 seeds per m-2) imposed to test the accuracy of predictions of grain N concentration in crops of varying canopy structure. Destructive samples were taken to determine total crop N content and canopy N distribution at ear emergence and harvest. Measurements of soil mineral N availability, ear numbers per m-2 and the number of spikelets per ear were made at ear emergence. Final grain yield and quality were also determined at harvest. Grain N concentration is the quotient of grain N content and grain yield. Both grain N content and yield explained a significant amount of the variation in grain N concentration observed across sites and fertiliser N treatments indicating that estimates of both must be included in models to predict N concentration. Grain N content was strongly related to total crop N content at harvest (P < 0.001; R2 = 0.96), which in turn was related to canopy N content at ear emergence (P < 0.001; R2 = 0.94). Similarly, grain yield was strongly related to total crop biomass at harvest (P < 0.001; R2 = 0.83), which in turn was related to crop biomass at ear emergence (P < 0.001; R2 = 0.88). These results indicated that predictions of grain N concentration might be possible from measurements of crop N content and biomass at ear emergence and that the effects of variation in harvest index, nitrogen harvest index and post-anthesis N uptake on grain N concentration are likely to be negligible under normal agronomic conditions in Ireland. Weather conditions in 2013 were unusually dry and estimates of soil moisture deficit and available water capacity indicated that the crop was water stressed. In 2014 weather conditions were close to the long term averages for the sites. Multiple regression models using canopy N content and biomass at ear emergence as explanatory variables accounted for 91% of the variation in grain N concentration when data from 2014 were used and 80% when data from both 2013 and 2014 were combined. The models developed using data from plots sown at 300 seed per m-2 in 2014 were tested against independent data from plots sown at 150 seeds per m-2 in the same year and at the same sites to test the accuracy of predictions across plant populations and canopy structures. The models were also tested using data from experimental plots and commercial fields collected in 2015 to test the accuracy of predictions in a different year across a range of sites and varieties. Values of grain N concentration predicted from measurements at ear emergence were compared with actual grain N concentrations measured at harvest. The accuracy of predictions was good with an R2 of 0.80 and RMSE of 0.114 %N for the test across seed rates and R2 of 0.80 and RMSE 0.220 %N for the validation in 2015 across sites and varieties. In 2014, grain N concentrations were increased significantly by applications of additional N fertiliser at anthesis with apparent recoveries (increase in N content (kg) /kg fertiliser N applied) in grain averaging 50% over the range of application rates indicating scope for increasing grain N concentration in crops predicted to be at risk of not meeting malting specifications Non-destructive measurements displayed significant relationships with N content and biomass at GS 59 across a combination of sites and seasons. However, issues in performance relating to instrument saturation were obvious and estimates never produced more accurate predictions of grain N concentration than destructive sampling. The results show that grain N concentration of spring barley can be predicted with good accuracy from measurements of canopy N and crop biomass made at ear emergence when the weather conditions are comparable to the long term average for the region. As conditions of drought are rare in Ireland, the prediction models are a potentially valuable tool to aid crop management and post-harvest operations by growers and maltsters. Further testing will be needed before users can be confident in the reliability of predictions over years and a larger set of varieties.

Samples of stores were drawn from two chains, Pizza Hut and Zale Corporation. Two different samples were taken from Pizza Hut. Site specific material and sales data were furnished by the companies and demographic material relative to each site was gathered. Analysis of variance tests for linearity were run on the three regression equations developed from the data and each of the three regressions equations were found to have a statistically significant linear relationship. Statistically significant differences were found among similar variables used in the prediction of sales by using Fisher's Z' Transformations on the correlation coefficients. Eight of the eighteen variables used in the Pizza Hut study were found to be statistically different between the two regions used in the study. Additionally, analysis of variance tests were used to show that traffic pattern variables were not better predictors than demographic variables.

O crescimento populacional e a degradação dos corpos d’água vêm exercendo pressão à agricultura moderna, a proporcionar respostas mais eficientes quanto ao uso racional da água. Para uma melhor utilização dos recursos hídricos, faz-se necessário compreender o movimento da água na natureza, onde o conhecimento prévio dos fenômenos atmosféricos constitui uma importante ferramenta no planejamento de atividades que utilizam os recursos hídricos como fonte primária de abastecimento. Nesse trabalho foram realizadas previsões de longo prazo com antecedência de sete meses e intervalo de tempo mensal de níveis no Sistema Hidrológico do Taim, utilizando previsões de precipitação geradas por um modelo de circulação global. Para realizar as previsões foi elaborado um modelo hidrológico empírico de regressão, onde foram utilizadas técnicas estatísticas de análise e manipulação de séries históricas para correlacionar os dados disponíveis aos níveis (volumes) de água no banhado. Partindo do pressuposto que as previsões meteorológicas são a maior fonte de incerteza na previsão hidrológica, foi utilizada a técnica de previsão por conjunto (ensemble) e dados do modelo COLA, com 30 membros, para quantificar as incertezas envolvidas. Foi elaborado um algoritmo para gerar todas as possibilidades de regressão linear múltipla com os dados disponíveis, onde oito equações candidatas foram selecionadas para realizar as previsões. Numa análise preliminar dos dados de entrada de precipitações previstas foi observado que o modelo de circulação global não representou os extremos observados de forma satisfatória, sendo executado um processo de remoção do viés. O modelo de empírico de simulação foi posteriormente executado em modo continuo, gerando previsões de longo prazo de níveis para os próximos sete meses, para cada mês no período de junho/2004 a dezembro/2011. Os resultados obtidos mostraram que a metodologia utilizada obteve bons resultados, com desempenho satisfatórios até o terceiro mês, decaindo seu desempenho nos meses posteriores, mas configurando-se em uma ferramenta para auxílio à gestão dos recursos hídricos do local de estudo. / Population growth and degradation of water bodies have been pressuring modern agriculture, to provide more efficient responses about the rational use of water. For a better use of water resources, it is necessary to understand the movement of water in nature, where prior knowledge of atmospheric phenomena is an important tool in planning activities that use water as the primary source of supply. In this study were performed long-term forecasts of water levels (seven months of horizon, monthly time-step) in the Hydrological System Taim, using rainfall forecasts generated by a global circulation model as input. To perform predictions was developed an empirical hydrological regression model. This model was developed based on statistical techniques of analysis and manipulation of historical data to correlate the input data available to the levels (volume) of water in a wetland. Assuming that weather forecasts are a major source of uncertainty in hydrological forecasting, we used an ensemble forecast from COLA 2.2 with 30 members to quantify the uncertainties involved. An algorithm was developed to generate all the multiple linear regression models with the available data, where eight candidates equations were selected for hydrological forecasting. In a preliminary analysis of the precipitation forecast was observed that the global circulation model did not achieve a good representation of extremes values, thus a process of bias removal was carried out. Then the empirical model was used to generate water levels forecast for the next seven months, in each month of the period june/2004 to december/2011. The results showed that the methodology used has a satisfactory performance until the lead time three (third month in the future) where the performance starts to show lower values. Beside the sharply lost of performance in the last lead times, the model is a support tool that can help the decision making in the management of water resources for the study case.

The objective of this study was to evaluate the breed additive genetic and heterozygotic effects, direct and maternal, on the weight and the average daily gain for a multibreed population Aberdeen Angus-Nellore. They were studied the characteristics adjusted weight to 205 days of age (P205) and to 550 days of age (P550) and average daily gain from birth to weaning (GMD_ND) and from weaning to 550 days of age (GMD_DS). The data was furnished by Gensys Consultores Associados S/C Ltda. and Natura Genética Sul Americana, and it contained records on 121,241 animals sired by 1,359 bulls and 84,465 cows, raised in 104 farms in differents brazilians states from 1986 to 2002. The data was analyzed by two different methodologies: Genetic Group Model (MGG) and Multiple Regression Method (MRM). In paper number one, pre-weaning period, it were used records on 96,502 animals, sired by 962 bulls and 67,231 cows, grouped in 2,542 contemporaneous groups (GC) and 45 genetic groups (GG). The model used for the analysis by MGG, included as fixed, the effects of contemporaneous group-GC (including the animals born at the same farm, year and season and belonging to the same sex and management group), genetic group-GG (genetic additive effect of the animal individual breed, paternal and maternal), and the covariables age of the animal at weaning (IDD) and age of the cow at parturition (IVP), as linear and quadratic effects. For the MRM analysis, the data was firstly adjusted for the effects of GC, and for the covariables IDD and IVP, linear and quadratic. After that, it was performed a MRM considering the direct and maternal additive genetic effect of the breed and the direct and maternal heterozigosity. In the paper number two, post-weaning period, it were evaluated performances of 40,310 animals, sired by 720 bulls and 31,136 cows, distributed in 650 contemporaneous groups and 33 genetic groups. The model adopted by MGG considered as fixed, the effects of contemporaneous group (including the animals born at the same farm, year and season and belonging to the same sex and management group at 550 days of age) and the covariables IVP, linear and quadratic effects and the age of the animal to the age standard of 550 days was done (ISA), as linear effect. The MRM analysis was performed at the same way that for paper one. For the pre-weaning period, the results of this work alloud us to conclude that the direct additive genetic effects from Aberdeen Angus breed in relation to Nellore breed give better performances; the direct and maternal heterozygosity effect were high and significant, suggesting that crossed calves and cows can perform better than the pure breeds. For the pre-weaning period the estimatives obtained by MRM were similar to those obtained by MGG. In relation to the post-weaning period it were observed differences in the performances of the animals, showing a better efficiency of Aberdeen Angus breed additive genes in relation to the Nellore breed. By the regression coefficient for direct and maternal effects it is possible to conclude that exist a bigger influence of the direct in relation to the maternal effect, suggesting that, in this phase, the animal performance depends most of its breed composition than of the influence of it mother. In spite of to exist similarities among the estimatives obtained by MGG and MRM, it is not recommended the use of MRM because of the inconsistencies observed in the results by this method. / O presente estudo teve por objetivo avaliar os efeitos genéticos aditivos raciais e heterozigóticos, diretos e maternos, sobre o desempenho ponderal de uma população multirracial Aberdeen Angus-Nelore. Foram estudadas as variáveis peso ajustado aos 205 dias de idade (P205), ganho médio diário do nascimento a desmama (GMD_ND), peso ajustado aos 550 dias de idade (P550) e ganho médio diário da desmama ao sobreano. O arquivo de dados foi fornecido pelas empresas Gensys Consultores Associados S/C Ltda. e Natura Genética Sul Americana, e continha informações de desempenho de 121.241 animais filhos de 1.359 touros e 84.465 vacas, criados em 104 fazendas, em diferentes estados brasileiros, coletadas no período entre 1986 e 2002. Os dados foram analisados através de dois modelos: Modelo de Grupo Genético (MGG) e Regressão Múltipla (MRM). No artigo 1, que abrangeu a fase prédesmama, foram utilizadas observações de 96.502 animais, filhos de 962 touros e de 67.231 vacas, agrupados em 2.542 grupos de contemporâneos e 45 grupos genéticos diferentes. O modelo utilizado para a análise através do MGG continha os efeitos fixos de grupo de contemporâneos (fazenda, ano de nascimento, sexo, estação de nascimento e grupo de manejo até a desmama) e grupo genético (efeito aditivo de raça individual, paterno e materno), além das covariáveis idade do bezerro a desmama e idade da vaca ao parto, incluídas como efeito linear e quadrático. Para o modelo de regressão múltipla, os dados foram inicialmente ajustados para o efeito GC, e das covariáveis idade do bezerro a desmama e idade da vaca ao parto, incluídas como efeito linear e quadrático, e então, foi conduzida a análise de regressão múltipla num modelo que considerou o efeito aditivo direto de raça, efeito aditivo materno de raça, heterozigose direta e heterozigose materna. No artigo 2, foi estudada a fase de pós-desmama, avaliando o desempenho de 40.310 animais, filhos de 720 touros e 31.136 vacas, distribuídos em 650 grupos de contemporâneos e 33 diferentes grupos genéticos, onde o modelo utilizado para o MGG continha os seguintes efeitos fixos de grupo de contemporâneos (fazenda, ano e estação de nascimento, sexo e grupo de manejo ao sobreano) e grupo genético, além das covariáveis a idade da vaca ao parto (efeito linear e quadrático) e idade do bezerro ao sobreano (efeito linear). A análise de regressão múltipla foi conduzida de forma similar ao artigo 1. Para a fase pré-desmama conclui-se que os efeitos aditivos diretos da raça Angus em relação à raça Nelore determinaram melhores desempenhos; os efeitos de heterozigose direta e materna foram altos e significativos, indicando um maior desempenho de vacas e bezerros cruzados. Para esta fase as estimativas geradas através de RM foram semelhantes àquelas encontradas através do MGG. Em relação a fase pós-desmama foi possível evidenciar que as diferenças observadas no desempenho dos animais, evidenciam a maior eficiência dos genes de efeito aditivo da raça Angus em contraste com os da raça Nelore e, através dos coeficientes de regressão para os efeitos diretos e maternos, é possível inferir que existe grande influência do efeito direto em relação ao materno, sugerindo que, nesta fase, o desempenho dos animais depende mais da sua própria composição genética racial do que da composição da mãe. Ainda que existam semelhanças entre os métodos de RM e MGG, para esta fase, não se aconselha a utilização da RM, visto da inconstância dos resultados obtidos através deste método.

Resumo: O objetivo deste trabalho foi investigar e modelar matematicamente a influência das variáveis do processo de têmpera e revenimento no limite de resistência à tração de arames de aço trefilado SAE 9254 com diâmetros de 2,00mm e 6,50mm. Para isso, utilizaram-se as ferramentas estatísticas de delineamento de experimentos e de regressão múltipla com aplicação de análise de componentes principais. Os resultados revelaram quais variáveis do processo têm influência significativa na variável resposta. Os modelos matemáticos obtidos pela regressão múltipla predizem com boa precisão a propriedade mecânica limite de resistência à tração. / Abstract: The purpose of this work was to investigate and to model mathematically the influence of the quench hardening and draw tempering process variables on the tensile strength of SAE 9254 drawn steel wires with 2,00 mm and 6,50mm diameter. To this end, the statistical tools of design of experiments and multiple regression with application of principal components analysis were employed. The results revealed the process variables that significantly affect the response variable. The mathematical models obtained by multiple regression were able to predict, with a reasonable accuracy, the mechanical property of the tensile strength. / Orientador: Messias Borges Silva / Coorientador: Rosinei Batista Ribeiro / Banca: Marco Antonio Carvalho Pereira / Banca: Gilbert Silva / Mestre

Made available in DSpace on 2014-06-11T19:30:10Z (GMT). No. of bitstreams: 0 Previous issue date: 2009-02-19Bitstream added on 2014-06-13T19:39:26Z : No. of bitstreams: 1 pimenta_cd_me_guara.pdf: 442019 bytes, checksum: 6f6f8b2d9fedc6b5b27318f6af0026ee (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / O objetivo deste trabalho foi investigar e modelar matematicamente a influência das variáveis do processo de têmpera e revenimento no limite de resistência à tração de arames de aço trefilado SAE 9254 com diâmetros de 2,00mm e 6,50mm. Para isso, utilizaram-se as ferramentas estatísticas de delineamento de experimentos e de regressão múltipla com aplicação de análise de componentes principais. Os resultados revelaram quais variáveis do processo têm influência significativa na variável resposta. Os modelos matemáticos obtidos pela regressão múltipla predizem com boa precisão a propriedade mecânica limite de resistência à tração. / The purpose of this work was to investigate and to model mathematically the influence of the quench hardening and draw tempering process variables on the tensile strength of SAE 9254 drawn steel wires with 2,00 mm and 6,50mm diameter. To this end, the statistical tools of design of experiments and multiple regression with application of principal components analysis were employed. The results revealed the process variables that significantly affect the response variable. The mathematical models obtained by multiple regression were able to predict, with a reasonable accuracy, the mechanical property of the tensile strength.

O crescimento populacional e a degradação dos corpos d’água vêm exercendo pressão à agricultura moderna, a proporcionar respostas mais eficientes quanto ao uso racional da água. Para uma melhor utilização dos recursos hídricos, faz-se necessário compreender o movimento da água na natureza, onde o conhecimento prévio dos fenômenos atmosféricos constitui uma importante ferramenta no planejamento de atividades que utilizam os recursos hídricos como fonte primária de abastecimento. Nesse trabalho foram realizadas previsões de longo prazo com antecedência de sete meses e intervalo de tempo mensal de níveis no Sistema Hidrológico do Taim, utilizando previsões de precipitação geradas por um modelo de circulação global. Para realizar as previsões foi elaborado um modelo hidrológico empírico de regressão, onde foram utilizadas técnicas estatísticas de análise e manipulação de séries históricas para correlacionar os dados disponíveis aos níveis (volumes) de água no banhado. Partindo do pressuposto que as previsões meteorológicas são a maior fonte de incerteza na previsão hidrológica, foi utilizada a técnica de previsão por conjunto (ensemble) e dados do modelo COLA, com 30 membros, para quantificar as incertezas envolvidas. Foi elaborado um algoritmo para gerar todas as possibilidades de regressão linear múltipla com os dados disponíveis, onde oito equações candidatas foram selecionadas para realizar as previsões. Numa análise preliminar dos dados de entrada de precipitações previstas foi observado que o modelo de circulação global não representou os extremos observados de forma satisfatória, sendo executado um processo de remoção do viés. O modelo de empírico de simulação foi posteriormente executado em modo continuo, gerando previsões de longo prazo de níveis para os próximos sete meses, para cada mês no período de junho/2004 a dezembro/2011. Os resultados obtidos mostraram que a metodologia utilizada obteve bons resultados, com desempenho satisfatórios até o terceiro mês, decaindo seu desempenho nos meses posteriores, mas configurando-se em uma ferramenta para auxílio à gestão dos recursos hídricos do local de estudo. / Population growth and degradation of water bodies have been pressuring modern agriculture, to provide more efficient responses about the rational use of water. For a better use of water resources, it is necessary to understand the movement of water in nature, where prior knowledge of atmospheric phenomena is an important tool in planning activities that use water as the primary source of supply. In this study were performed long-term forecasts of water levels (seven months of horizon, monthly time-step) in the Hydrological System Taim, using rainfall forecasts generated by a global circulation model as input. To perform predictions was developed an empirical hydrological regression model. This model was developed based on statistical techniques of analysis and manipulation of historical data to correlate the input data available to the levels (volume) of water in a wetland. Assuming that weather forecasts are a major source of uncertainty in hydrological forecasting, we used an ensemble forecast from COLA 2.2 with 30 members to quantify the uncertainties involved. An algorithm was developed to generate all the multiple linear regression models with the available data, where eight candidates equations were selected for hydrological forecasting. In a preliminary analysis of the precipitation forecast was observed that the global circulation model did not achieve a good representation of extremes values, thus a process of bias removal was carried out. Then the empirical model was used to generate water levels forecast for the next seven months, in each month of the period june/2004 to december/2011. The results showed that the methodology used has a satisfactory performance until the lead time three (third month in the future) where the performance starts to show lower values. Beside the sharply lost of performance in the last lead times, the model is a support tool that can help the decision making in the management of water resources for the study case.

nÃo hà / A inovaÃÃo à um processo dinÃmico que possibilita a implementaÃÃo de algo novo ou substancialmente aperfeiÃoado, podendo suscitar mudanÃas tanto na empresa quanto no contexto socioeconÃmico o qual està inserido. Os resultados inovativos sÃo diversos, uma vez que sua prÃpria criaÃÃo pode encadear novos processos ou melhorias nos procedimentos e tÃcnicas empregados na produÃÃo ou no fornecimento de serviÃos. A proposta deste trabalho à analisar quais os fatores determinantes da implementaÃÃo de inovaÃÃes de produto e/ou processo pelas empresas brasileiras. Para este fim, se utilizou dados provenientes da Pesquisa de InovaÃÃo (PINTEC 2011), realizada pelo Instituto Brasileiro de Geografia e EstatÃstica (IBGE), com apoio do MCTI e FINEP. Os dados tÃm como referÃncia o total Brasil de empresas do setor industrial e de serviÃo, no perÃodo de 2009 a 2011. Com base na literatura, construiu-se um modelo economÃtrico relacionando fatores que podem influenciar o percentual de empresas que implementam inovaÃÃes. Algumas hipÃteses foram levantadas e testadas empiricamente. O modelo tem como variÃveis independentes o apoio do governo, relaÃÃes de cooperaÃÃo com outras organizaÃÃes, dispÃndio em atividades inovativas sobre a receita lÃquida de vendas, pessoas ocupadas em atividades internas de P&D com nÃvel superior e financiamento por recursos de terceiros, mensuradas proporcionalmente ao nÃmero de empresas investigadas. Os resultados obtidos confirmam a hipÃtese formulada de que o apoio do governo està positivamente relacionado ao percentual de empresas brasileiras que implementam inovaÃÃes de produto e/ou processo. E, no caso de inovaÃÃo de processo, dispÃndios em atividades inovativas e qualificaÃÃo profissional tambÃm se mostraram como fatores significativos. A significÃncia global do modelo confirmou a atuaÃÃo conjunta dos fatores como determinantes da implementaÃÃo de inovaÃÃes de produtos e/ou processos novos ou substancialmente aprimorados. / Innovation is a dynamic process that enables the implementation of something new or substantially improved, changes may raise both the company and the socioeconomic context which is inserted. Innovative results are various, since his own creation can chain new processes or improvements to procedures and techniques employed in the production or supply of services. The purpose of this project is to analyze the determinants of implementation of innovations in product or process by braziliancompanies. For this purpose, used data from the Research of Innovation ( PINTEC 2011), conducted by the Instituto Brasileiro de Geografia e EstatÃstica (IBGE), with support from FINEP and MCTI. The data has reference to the total Brazil, the companies in the industrial and service sector in the period 2009-2011. Based on the literature, constructed an econometric model relating factors that can influence the percentage of companies that implement innovations. Some hypotheses were developed and tested empirically. The model has as independent variables the government's support, cooperative relationships with other organizations, expenditure on innovation activities of the sales revenue, persons employed in internal R&D with degree and funding from third resources, measured in proportion to the number companies investigated. The results confirm the hypothesis formulated that government support is positively related to the percentage of brazilian companies that implement product innovations and/or process. And in the case of process innovation, expenditure on innovation activities and professional qualification also proved to be significant factors. The global significance of the model confirmed the joint action of the factors as determinants of the implementation of product innovations and/or new or significantly improved processes.