Key Controlled Variable Model of MSWI Process Based on Ensembled Decision Tree Algorithm

Incineration is one of the main techniques for municipal solid waste (MSW) treatment, which is widely used in most countries. The stable operation of the MSW incineration (MSWI) process depends on three key controlled variables, namely furnace temperature, flue gas oxygen concentration, and boiler steam flow. The first problem to develop the intelligent control and optimization technology of the MSWI process is to establish the controlled object model. Thus, a data-driven model based on ensemble decision tree algorithm is constructed for the above-mentioned three controlled objects in this paper. It is an ensemble combination of random forest (RF) and gradient boosting decision tree (GBDT). First, the random sampling strategy is used to pre-process a large amount of MSWI process data to obtain modeling sub datasets. Second, the training subsets are used to build the RF model. Furthermore, a GBDT model with a serial structure is constructed by using the gradient iterative RF model. Finally, the predicted values are produced by the simple weighted average of the predicted output of the RF and GBDT sub-model. The results show the effectiveness of the proposed method on the basis of the real-time process data obtained in the actual MSWI process.