DETECTION OF KNOCK IN SPARK-IGNTION ENGINE: VIBRATION PATTERN FORudDETERMINE SPARK TIMING INGITION USINGudFUZZY CONTROL AND ARTIFICIAL NEURAL NETWORK

摘要

The problem of detonation (knocking) in the internal combuction engines, especially inudthe Otto (petrol) engine, that makes some damages, low fuel economy and performance. Untiludnow, the detonation problem has not been solved completely. In this paper, the problem hadudbeen studied to discover the new method, how to reject detonation without any changes ofudengine structure, easy to install and can be activate immediately.udThe detonation phenomenon have been explained by three theory, that are autoudignition theory, detonation theory and flame vibration, that make noise, detonation (knocking)udand damages of some equipments. The detonation can be cause by many things, such as highudcompression ratio, low grade of fuel, bad combustion camber, low turbulence, large sparkudadvance (timing).udFor the conventional ignition system, a governor and a vacuum advancer control theudignition timing. It is reliable mechanism, but can not work properly at all conditions. Most ofudthem make detonation occur at low speed and low endurance of the contact breaker. The lastudtechnology, electronically device with detonation sensor replace the conventional system.udCombination ignition system, the new system, integration of conventional and electronicallyudsystem, can control the detonation, by making correction of the spark advance (ignition timing)udautomatically in all condition of speed and load (throttle), using fuzzy logic reasoningud(calculation) in the microcontroller. This research also tries to detect detonation by using theudvibration pattern with the hope of detonation detection by this method will be more stable andudreliable. Detonation detection with machine vibration recordings was be done by measuring itsudintensity, but this often happens is unstable.udThe result of using the new (combination) ignition system in the engine is the rejecteduddetonation, improving performance and efficiency. Another result, the motor vibration patternudcan be recorded and the vibration pattern of detonation occurs at high frequency. It has audcertain pattern in synch with his ignition. For the further the research will use the Neuro-FuzzyudAdaptive Control for more applicable in the engine.