The thesis focuses on a hydraulic system design for a specific application of hydrostatic drive with regard to energy efficiency of proposed solution. Standard hydraulic and electronic components are used, mostly products of Bosch Rexroth. The aim of the thesis is to carry out a complex design from mechanic through hydraulic up to the electronic part of the system. After a short introduction in the beginning of the thesis is the particular application described and analyzed. Then the elemental principles of hydraulic devices and systems are explained. In the following part the actual solution is presented, which expands to several different alternatives. These alternatives are compared according to set criteria and the selected alternative is then further developed. At the end of the thesis, the outputs are evaluated and the direction of further possible improvements in the design is outlined.

In this thesis study, a valveless hydraulic servo system controlled by two pumps is investigated and its performance characteristics are compared with a conventional valve controlled system both experimentally and analytically. The two control techniques are applied on the position control of a single rod linear actuator. In the valve controlled system, the flow rate through the actuator is regulated with a servovalve / whereas in the pump controlled system, two variable speed pumps driven by servomotors regulate the flow rate according to the needs of the system, thus eliminating the valve losses. To understand the dynamic behaviors of two systems, the order of the differential equations defining the system dynamics of the both systems are reduced by using the fact that the dynamic pressure changes in the hydraulic cylinder chambers become linearly dependent on leakage coefficients and cylinder chamber volumes above and below some prescribed cut off frequencies. Thus the open loop speed response of the pump controlled and valve controlled systems are defined by v second order transfer functions. The two systems are modeled in MATLAB Simulink environment and the assumptions are validated. For the position control of the single rod hydraulic actuator, a linear state feedback control scheme is applied. Its state feedback gains are determined by using the linear and linearized reduced order dynamic system equations. A linear Kalman filter for pump controlled system and an unscented Kalman filter for valve controlled system are designed for estimation and filtering purposes. The dynamic performances of both systems are investigated on an experimental test set up developed by conducting open loop and closed loop frequency response and step response tests. MATLAB Real Time Windows Target (RTWT) module is used in the tests for application purposes.

Submitted by Cláudia Bueno (claudiamoura18@gmail.com) on 2016-01-29T17:14:46Z No. of bitstreams: 2 Dissertação - Ricardo Santos Coutinho - 2015.pdf: 10137543 bytes, checksum: 50b06893fd31466aedf36ff57ca15495 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2016-02-01T11:58:11Z (GMT) No. of bitstreams: 2 Dissertação - Ricardo Santos Coutinho - 2015.pdf: 10137543 bytes, checksum: 50b06893fd31466aedf36ff57ca15495 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2016-02-01T11:58:11Z (GMT). No. of bitstreams: 2 Dissertação - Ricardo Santos Coutinho - 2015.pdf: 10137543 bytes, checksum: 50b06893fd31466aedf36ff57ca15495 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2015-12-10 / In water distribution systems, most of the electricity is consumed by pumping systems. Thus, reducing the inefficiency of the pump can provide significant energy savings. This can be achieved by substituting fixed-speed pumps by variable-speed pumps (VSP). However, the high costs of installation and operation of the VSPs require a preliminary economic assessment of the design, which is usually accompanied by the use of hydraulic solvers to simulate the behavior of the water network. One of the most used softwares for this purpose is the EPANET 2, which uses the gradient method for the hydraulic analysis of the steady flow. However, the EPANET 2 is inaccurate when computing the VSP efficiency. This is because the software does not take into account the changes in the pump efficiency curve when the pump speed is changed. In this study, a routine for calculating VSPs efficiency, based on the Sârbu and Borza (1998) formula, is implemented in the source code of the EPANET 2. The influence of the new routine on the results predicted by the software for efficiency, power and energy consumption of the VSP is evaluated in simulations with a hypothetical network. The results showed that the original version of EPANET 2 tends to underestimate the efficiency and thus overestimate the power and energy consumption of the VSPs. In the present case, the energy consumption of the system predicted by the original version of EPANET 2 was 21.99% higher than the predicted by the modified version. These results point out that the modified version of EPANET 2 provides a better prediction of VSPs behavior and thus its use is recommended in the simulation of water distribution networks equipped with VSPs. / Em redes de distribuição de água, a maior parte da energia elétrica é consumida pelos sistemas de bombeamento. Deste modo, a redução da ineficiência das bombas pode proporcionar uma economia expressiva de energia. Isto pode ser alcançado pela substituição de bombas de rotação constante por bombas de rotação variável (BRVs). No entanto, os elevados custos de instalação e operação de BRVs requerem uma avaliação econômica prévia do projeto, a qual geralmente é acompanhada do uso de simuladores hidráulicos para simular o comportamento da rede. Um dos programas mais utilizados para este fim é o EPANET 2, o qual emprega o Método Gradiente para a análise hidráulica do escoamento permanente. Todavia, o EPANET 2 é impreciso ao computar a eficiência de BRVs. Isto ocorre porque o programa não leva em conta a mudança na curva de eficiência da bomba quando sua velocidade de rotação é alterada. No presente trabalho, uma rotina para o cálculo da eficiência de BRVs, baseada na fórmula de Sârbu e Borza (1998), é implementada no código-fonte do EPANET 2. A influência da nova rotina nos resultados previstos pelo programa para a eficiência, potência e consumo de energia de BRVs é avaliada em simulações com uma rede hipotética. Os resultados mostraram que a versão original do EPANET 2 tende a subestimar a eficiência e, consequentemente, superestimar a potência e o consumo de energia de BRVs. No caso em tela, o consumo de energia do sistema previsto pela versão original do EPANET 2 foi 21,99% maior que o previsto pela versão modificada. Tais resultados apontam que a versão modificada do EPANET 2 proporciona uma melhor previsão do comportamento de BRVs e, portanto, o seu uso é recomendado na simulação de redes de distribuição de água equipadas com BRVs.