Symbol-sequence statistics-based cylinder-to-cylinder variation control in spark-ignition engines

With the rapid market expansion of hybrid/electric vehicles, automobiles equipped with conventional internal combustion engines are facing stricter emission regulations and growing demands for efficiency improvement. As the bottleneck for further combustion efficiency enhancement, combustion variation in spark ignition engines has been investigated by means of both experimental tests and numerical analyses for many years. In this study, a symbol-sequence statistics-based method was utilized to detect and extract the deterministic features in combustion variation for a multi-cylinder engine. Based on the resulting data, a novel control algorithm was proposed to reduce cylinder-to-cylinder variation. The effectiveness and performance of the proposed method were experimentally validated on a production spark ignition engine test bench.