Power Plant Boiler Slagging Characteristics Prediction Based on Fuzzy C-Means Clustering and Support Vector Machine

The support vector machine based on fuzzy C-means clustering algorithm is used to predict the boiler slagging characteristics. The model regards the complex training sample set as multiple mixed groups, and establishes a single SVM model for each group, which makes the prediction problem of large-scale data set become a multi-model prediction problem. After pre-processing, the SVM algorithm retains the sample points near the optimal classification hyperplane, removes the distant sample points, reduces the overfitting degree of the training sample, and improves the prediction accuracy. The input variables of the model are the softening temperature t<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">ST,</inf> the alkali-acid ratio B/A, the silica-aluminum ratio w(SiO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>)/w(Al<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>O<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3</inf>) and the silicon ratio G, as well as the dimensionless furnace average temperature φ<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">t</inf> and the dimensionless tangential circle diameter φ<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">d</inf>, and the output variable is the slagging degree. In this paper, 40 groups of boiler slagging samples are used, of which 30 groups of data are used as training samples to train fuzzy C-means clustering support vector machine model parameters, and 10 groups of data are used as test samples to verify the performance of the model. The optimized model is used to evaluate the slagging characteristics of 10 boilers, and the accuracy of the prediction results is 100%. The FCM-SVM model can achieve accurate prediction of the slagging characteristics of boilers.