【作者】

Francisco Jimenez Espadafor

【摘要】

该论文已在芬兰赫尔辛基举行的第28届CIMAC大会上发表。论文的版权归CIMAC所有。High efficiency, low cost of the fuel and long lifespan are the guaranty for low speed diesel engines to be the solution in many application areas like power generation or marine propulsion. On the other side low speed diesel engines show reduced availability when compared to steam turbines or, in a less grade, to gas turbines. The state of mechanical parts as piston, cylinders, rings, valves or injectors is solved usually through the operational maintenance, where detailed knowledge of turbocharger vibrations is not usually at disposal. This paper shows the performance and main features of a diagnosis system currently in service on four 14 MW-125 rpm ten cylinders, two stroke uniflow-scavenged diesel generators. The system captures and processes the information provided by 17 sensors: one accelerometer per cylinder head, one accelerometer per each turbocharger, one pick-up for each turbocharger for rpm measurement, one pick-up for instantaneous engine speed at fly wheel and one instantaneous torque sensor formed by strange gages located in the shaft between engine and generator. The diagnosis system is based on data-driven techniques supported by the historical data of the diesel generator at good condition. For the engine, the system captures about one thousands of engine cycle per day and applies windowing to every relevant engine process which allows the supervision of the following events: Engine: needle injector opening, needle injector closing, start of injection angle, end of injection angle, combustion roughness, compression stroke, expansion stroke, inlet port opening, exhaust valve closing, indicated torque harmonics. Generator: harmonics of the generator rotor oscillation. It has been demonstrated that harmonics 24th, 48th, 72th and 96th are related to the transitory torque generated by the passing of the rotor armature in front of the stator and therefore to its technical condition. For the turbochargers, the analysis is performed for a package of 80 consecutive turbocharger cycles which allows the supervision of the following events: Turbocharger: Engine Power-Turbocharger rpm curve, bearing vibration harmonics, turbocharger shaft vibration harmonics. Each event is characterized by an indicator (harmonic amplitude, root mean square acceleration amplitude at specific frequency band, torque and angular speed amplitude at specific frequency band, mean turbocharger speed). The reference condition is evaluated from the event indicator-power data map at good condition which is gathered after engine overhaul or after the replacement of the component. The condition of the engine generator is evaluated from the evolution of the symptom named SINDEX along operational hours of the generator: SINDEX = 2 • abs[Ic (W) - It max (W)] / [It max (W) + It min (W)] if Ic (W) > It max (W) SINDEX = 2 • abs[Ic (W) - It min (W)] / [It max (W) + It min (W)] if Ic (W) < It min (W) Being I the event indicator, W generator power, suffix c the current value, suffix t the reference template value and max and min refers to the maximum and minimum values of the indicators template for every generator power band. The system is working from 2012 in one group and from 2015 in the three identical diesel groups at Diesel Power Plant of Mahón (Balearic Island, Spain). Up to now the system has identified the following anomalies: • Cylinder seizure (3) • Broken rings (9) • Injector malfunction (5) • Changing of the start of injection (3) • Exhaust valve regulator failure (2) • Turbine blade failure (1) • Compressor fouling (continuously) Work is in progress to increase the capability of the system to identify combustion anomalies that will be developed before 28th CIMAC Congress.

【会议名称】

第28届CIMAC会议

【会议地点】

芬兰 赫尔辛基

【下载次数】

2

返回