A New Fault Diagnosis Model for Circuits in Railway Vehicle Based on the Principal Component Analysis and the Belief Rule Base

Table 1

PCA-BRB fault diagnosis model for circuits.

---

Step 1: fault feature selection by the PCA method.

Step 1.1: constructing the standardized data by equations (1)–(3).

Step 1.2: calculating the correlation matrix R by equation (4) and calculating the eigenvalue and characteristic vector of R matrix.

Step 1.3: calculating the principal component of eigenvalue by equations (6) and (7).

Step 2: the establishment of the initial BRB.

Step 2.1: defining semantic values of system input and output.

Step 2.2: setting the initial parameters of the BRB.

Step 2.3: calculating the initial BRB output semantic value by equations (8)–(14).

Step 3: optimization of the BRB parameters.

Step 3.1: calculating the initial BRB and real output variance , as the target function by equation (20).

Step 3.2: establishing the constraint by equations (16)–(19).

Step 3.3: selecting the best parameters in the current generation.

Step 4: electronic circuit fault diagnosis.

Step 4.1: the testing data are transformed to the belief degrees by equation (11).

Step 4.2: calculating the activated weights of rules in the optimized BRB by equation (9) or (10).

Step 4.3: deriving the distributed output of the system's state by equations (13) and (14).

Step 4.4: calculating the expected utility of the fault diagnosis by equation (15).

---