Dynamic model based optimization control for combustion process of coke oven

In view of large range change of operating status of combustion process and difficulties in building global model and timely optimizing control parameters, this paper presents dynamic model based optimization control method using the property that just-in-time learning model can rapidly reflect process change characteristics. Based on the idea of local modeling, distance and change trend angle between samples are both considered to choose data and create an online sample library. With this dynamic sample library, a dynamic model for combustion process of coke oven is build. And based on the dynamic process model, the differential evolution algorithm is applied to optimization the universe parameters of fuzzy control to ensure the controller can meet the dynamic changes of the combustion process of coke oven. Simulations verified the effectiveness of the method.