No abstract

Mechanical problems with large-scale mining machinery can cause serious losses in operational efficiency and profits. A diesel engine with a misfiring cylinder is likely to break down in a short period of time when the engine is used without maintenance. The authors have developed a new system to detect whether an engine is misfiring and, if so, which cylinder is misfiring. This system will be able to manage the engine from a remote location. In this study, a minimum number of sensors will be applied on two cylinders of the engine in order to detect misfiring cylinder. The measured waveform was analyzed by the statistical analysis of R.M.S value. Dividing up combustion time of each cylinder is the first step for statistical analysis of R.M.S value. In case of idling state of the engine, it is easy to divide up combustion time. But, it is necessary to find engine revolutions under the state of rotational fluctuation of the engine. The period of combustion time vary every hour under the state of rotational fluctuation. Therefore, engine revolutions were computed by using Prony method. The period of combustion time was divided, and R.M.S value was calculated.

Mechanical problems, common in large mining equipment, result in serious losses in operations efficiency and financial return. In this study, a minimum number of sensors was applied on the side of two cylinders of a diesel engine in order to build an abnormality diagnostic system that detects anomalous engine behaviour at an early stage. The study investigates whether or not the misfiring of an engine cylinder can be detected through analysis of acceleration results using an aggregative learning method (ALM). The results show that, using ALM, distinctive differences could be observed in almost every cylinder except those farthest from the accelerometer.