An improved multi-source based soft sensor for measuring cement free lime content

The excess free lime (f-CaO) content in clinker is the main cause of cement instability. Thus, it is crucial to effectively measure f-CaO content in real time for implementing a closed-loop control. To improve the estimation accuracy and sensor's reliability, improved multi-source modeling techniques are developed in this paper. In this work, fuzzy entropy is employed to compress the feature vectors of a segmental point dataset to enhance the sensor model's generalization power, and a decorrelated neural-net ensemble (DNNE) with random weights is employed to build the soft sensor. In this way, experiments with comprehensive comparisons can be carried out. Experimental results indicate that the improved soft sensor performs favorably in terms of both prediction accuracy and model reliability, compared with other soft sensor models.

[1]  Tommy W. S. Chow,et al.  Comments on "Stochastic choice of basis functions in adaptive function approximation and the functional-link net" [and reply] , 1997, IEEE Trans. Neural Networks.

[2]  Ian H. Witten,et al.  The zero-frequency problem: Estimating the probabilities of novel events in adaptive text compression , 1991, IEEE Trans. Inf. Theory.

[3]  Tianyou Chai,et al.  Designing compact Gabor filter banks for efficient texture feature extraction , 2010, 2010 11th International Conference on Control Automation Robotics & Vision.

[4]  Y. Takefuji,et al.  Functional-link net computing: theory, system architecture, and functionalities , 1992, Computer.

[5]  B. Bakshi,et al.  On-line multiscale filtering of random and gross errors without process models , 1999 .

[6]  Carl Duchesne,et al.  Multivariate Image Analysis of Flames for Product Quality and Combustion Control in Rotary Kilns , 2006 .

[7]  Shankar Narasimhan,et al.  A theoretically rigorous approach to soft sensor development using Principal Components Analysis , 2011 .

[8]  Tianyou Chai,et al.  Flame Image-Based Burning State Recognition for Sintering Process of Rotary Kiln Using Heterogeneous Features and Fuzzy Integral , 2012, IEEE Transactions on Industrial Informatics.

[9]  Wangxin Yu,et al.  Characterization of Surface EMG Signal Based on Fuzzy Entropy , 2007, IEEE Transactions on Neural Systems and Rehabilitation Engineering.

[10]  Sten Bay Jørgensen,et al.  Soft sensor design by multivariate fusion of image features and process measurements , 2011 .

[11]  G LoweDavid,et al.  Distinctive Image Features from Scale-Invariant Keypoints , 2004 .

[12]  Dianhui Wang,et al.  Fast decorrelated neural network ensembles with random weights , 2014, Inf. Sci..

[13]  Yoh-Han Pao,et al.  Stochastic choice of basis functions in adaptive function approximation and the functional-link net , 1995, IEEE Trans. Neural Networks.

[14]  Tianyou Chai,et al.  Multisource Data Ensemble Modeling for Clinker Free Lime Content Estimate in Rotary Kiln Sintering Processes , 2015, IEEE Transactions on Systems, Man, and Cybernetics: Systems.

[15]  T. Landauer,et al.  Indexing by Latent Semantic Analysis , 1990 .

[16]  Sten Bay Jørgensen,et al.  A systematic approach for soft sensor development , 2007, Comput. Chem. Eng..

[17]  Alejandro F. Frangi,et al.  Two-dimensional PCA: a new approach to appearance-based face representation and recognition , 2004 .

[18]  Honglu Yu,et al.  Monitoring flames in an industrial boiler using multivariate image analysis , 2004 .

[19]  Peng Sun,et al.  Rotary kiln flame image segmentation based on FCM and gabor wavelet based texture coarseness , 2008, 2008 7th World Congress on Intelligent Control and Automation.

[20]  Bogdan Gabrys,et al.  Review of adaptation mechanisms for data-driven soft sensors , 2011, Comput. Chem. Eng..

[21]  Hare Krishna Mohanta,et al.  Development and comparison of neural network based soft sensors for online estimation of cement clinker quality. , 2013, ISA transactions.