Study On Fault Diagnosis Of Gasoline/CNG Bi-fuel Vehicle Based On Wavelet Analysis Method

The Gasoline/CNG Bi-fuel Vehicles have a huge proportion of Natural Gas Vehicles in our country. And they are mainly in urban public transport and taxis. Compared to the gasoline engine, the bi-fuel engine is not modified radically. It is reconstructed with a suit of gas equipment and the gas equipment is made from fuel supply system (mechanical parts) and controlling system (electronic controlling parts). The bi-fuel engine can not take advantage of the characteristic of superior octane number, low calorific value, and emission. By contraries, it has many problems, such as being deficient in power, the engine wearing with corrosion, being difficult in engine cold start, the fluky controlling system and so on. These problems and resolving method will be expatiated in this dissertation.The testing system which can be used in testing bi-fuel vehicle is established. In this system, gasoline and natural gas can be switched from one to another conveniently and the air-fuel ratio is in close-loop control. The oxygen sensor plays an important part in realizing close-loop control of air-fuel ratio. In the process of close-loop control, the electronic controlled unit modifies the injection impulse width or stepper motor by the signal of the oxygen sensor. The reason for losing efficacy of the oxygen sensor is discussed and some faults of it are simulated in the test. The characteristic of the oxygen sensor, the engine's behavior because of its faults and the characteristic of electronic controlled unit as to distinguishing faults are achieved in the test.The wavelet analysis method is the essential way in diagnosing fault of the electronic controlled engine. The normal and abnormal signal of sensors can be collected by oscillograph and the parameter of signal can be analyzed by oscillograph. In this way, the characteristic of the normal and abnormal signal can be found out. This is significative for diagnosis. The feasibility whether the wavelet analysis method can be applied to fault diagnosis of the electronic controlled engine or not is discussed in this dissertation. The normal and abnormal signal of the oxygen sensor is collected by oscillograph. The signal can be analyzed and de-noised by wavelet transform and the normal and abnormal signal can be distinguished easily. The normal signal of velocity sensor and the abnormal signal of it when some cylinders are off are collected. The wavelet coefficients are achieved by wavelet transform. These wavelet coefficients can show differences between normal and abnormal signal and can decide which cylinder breaks. It proves that using the wavelet analysis method to diagnose fault of the electronic controlled engine is feasible.