An improved combination modeling method for free lime content in cement clinker
Pengcheng Zhao, Zhibiao Zhao, Ying Chen · 2020
The free lime (fCaO) content in cement clinker is an important quality index of cement production process and main cause of cement instability. Aiming at the difficulty in real-time measuring of fCaO content, building the fCaO content model is particularly important. In this work, the ameliorated quantum- behaved particle swarm optimization (A-QPSO) algorithm, which local search area is modified with the number of iterations, is proposed based on using the sequential quadratic programming algorithm to enhance the local search ability. The A-QPSO is adopted to adjust the parameters of multiple kernel extreme learning machine. The improved combination modeling method is also employed to build the fCaO content model. Compared with other methods, the method proposed in this research performs favorably in terms of both regression accuracy and generalization ability according to the experimental results.