Fault Diagnosis Technology of Data-driven Industrial Process - Method Based on Principal Component Analysis and Partial Least Squares / Books of ... Study and Application (Chinese Edition)

In the book Fault Diagnosis Technology of Data-driven Industrial Process - Method Based on Principal Component Analysis and Partial Least Squares, the author expounds comprehensively various dynamic systems fault diagnosis technology development status and future trends in Chapter 1. From Chapter 2 to Chapter 6, the author introduces main element analysis model, and the model-based fault detection, isolation and identification methods. In this section the author also discusses various improvement methods for principal component analysis model, such as selection of number of principal components, dynamic time warping-based improvement, and non-normal-sub domain-based application. From Chapter 7 to Chapter 11, the author introduces partial least squares model, the model-based fault detection, reconstruction and diagnostic algorithms. This section contains the authors latest research achievement, namely the geometric interpretation of partial least squares model structure, as well as improvements for output related fault model. These studies achievement also reveals the fundamental difference and intrinsic link of the principal component analysis and partial least squares in the monitoring process. In Chapter 12 and Chapter 13, the author focuses on discussion of the latest research direction in the above-mentioned field - fault prediction issues of continuous multivariable process. Based on principal component analysis model and partial least squares model, the fault prediction issues are to be studied further.