Thesis (PH.D.) - Cairo University - Faculty Of Engineering - Department Of Aerospace Engineering

This research presents seventeen neural network approaches for faults diagnosis of hydraulic control systemsThree approaches for each of double acting hydraulic cylinder in position control system , hydraulic variable displacement swash plate axial piston pump drives a certain load , hydraulic motor in a closed loop position control system , relief valve in hydraulic circuit , and flow control valve in hydraulic circuit , and then two approaches for a complete hydraulic circuitThe double acting hydraulic cylinder operates in a closed loop position control system , the cylinder is controlled by electrically operated servo valve and moves a 50 - kg trolley mounted massThe trolley position is measured with a displacement transducer and this signal is used to close the position control loopRandom step input signals with small amplitudes are used to the demand transducerThe control system governing equations are deduced and solved numerically to determine the system dynamic behavior according to the demand random signalsThe attained behavior used as basis to diagnose faults for three main parameters , which are the system supply pressure , the internal leakage and the coefficient of friction of the hydraulic cylinderThe neural network is developed to diagnose single fault and two faults occurring simultaneously

Issued also as CD