Engine knock detection for a multifuel engine using engine block vibration with statistical approach

Engine knock is an obstacle for maximizing CNG fuel utilization on a Diesel-CNG Dual Fuel engine. Prolong experience of this phenomenon may lead to severe engine damage. The low intensity of this phenomenon is difficult to recognize due to other noises from the engine. Thus, improper engine tuning techniques may make this phenomenon unnoticeable until the engine damages. Knock phenomena on such engines may not be detected in the combustion analysis graph. Its random occurrence in consecutive engine cycles makes it difficult to be seen using visual data. This knowledge gap, if closed, can lead to an opportunity for knock avoidance on the multifuel engine. This work proposed a method to quantify the knock occurrence based on engine block vibration using a single piezoelectric knock sensor. The knock occurrence was detected by comparing the calculated knock index with the knock threshold, determined using a statistical three-sigma rule analysis. This method can index the knock intensity, detect the engine knock occurrence, and visualize the knock phenomenon. • This paper describes an alternative engine knock detection technique based on engine block vibration. • This method proposes the knock threshold determination based on statistical three-sigma rule analysis. • This method is capable of visualizing the knock phenomenon in consecutive and at each engine cycle.