Abstract
Modeling and simulation have become essential tools in fluid power engineering. On the overall system simulation level, most components are described by a concentrated parameter approach resulting in a system of ordinary differential equations. The exception to this rule is found in pipelines and hoses, where wave propagation may be of importance and hence a distributed parameter approach results in some sort of spatial discretization in the modeling process. For long bores in hydraulic valve manifolds and also for rigidly supported steel pipes, the dynamic behaviour can be predicted very well with standard models available from literature like the classical frequency dependent friciton model of D’Souza and Oldenburger. With wire or fibre reinforced high pressure rubber hoses, this is no longer true and the transfer behaviour of such transmission lines can only be characterized based on measurement data. The present paper presents a novel approach towards the construction of low order models for the approximation of the measured input-output behaviour of fluid power hose lines. The method is based on the measurement of the linearized system dynamics at a number of pressure levels and is intended towards the local linear model tree (LOLIMOT) technique after identifying reduced order models for each pressure level. The results are applicable to fluid power system simulations where the dynamics of hose lines becomes important due to high operating frequencies or when noise attenuation in the hose lines is the primary focus of the simulation. The measurements provided in the paper are for a fluid power hose with a diameter of 20 mm and a length of 3 m.
