Feedback error learning neural networks for spark advance control using cylinder pressure

This paper presents a spark advance control strategy based on the location of peak pressure (LPP) in spark ignition engines using artificial neural ne tworks. The well-known problems of the LPP-based spark advance control meth od are that many samples of data are required and there is a problem detect ing the combustion phasing owing to hook-back during lean burn operation, I n order to solve these problems, a feedforward multilayer perceptron networ k (MLPN) is introduced in this study. The LPP and hook-back are estimated u sing the MLPN, which needs only five samples of output voltage from the cyl inder pressure sensor. The estimated LPP can be regarded as an index for co mbustion phasing and can also be used as a minimum spark advance for best t orque (MBT) control parameter. The performance of the spark advance control ler is improved by adding a feedforward controller which reflects the abrup t changes of the engine operating conditions such as engine speed and manif old absolute pressure. The feedforward controller consists of the radial ba sis function network, and the feedback error learning method is used for th e training of the network. In addition, the proposed control algorithm does not need sensor calibration and pegging (bias calculation) procedures beca use the MLPN estimates the LPP from the raw sensor output voltage, The feas ibility of this methodology to control spark advances is closely examined t hrough steady and transient engine operations. The experimental results hav e revealed that the LPP shows favourable agreement with the optimal value e ven during the transient operation of the engine.