Mingming Wang

Improved probabilistic neural network PNN and its application to defect recognition in rock bolts

The probabilistic neural network (PNN) model is developed based on Bayesian criteria of multivariate pattern classification. Because PNN has the advantages of low training complexity, high stability, quick convergence, and simple construction, it can be well applied in the area of defect recognition. The application of PNN has a bottleneck problem; that is, the selection of smoothing parameters will seriously affect PNN recognition accuracy. In this paper, we propose an improved PNN model that employs a differential evolution algorithm to optimize the smoothing parameters. Furthermore, a defect recognition approach is proposed based on the improved PNN, which is successfully applied to an anchoring testing field. The proposed approach includes two key steps. The first step involves extracting the energy eigenvector of the defect signal with normalization based on a wavelet packet. The second step involves recognizing defects based on the improved PNN. Simulation results show that our improved PNN model is superior to the traditional PNN model. Thus, the improved PNN can provide useful references for recognizing the defect type of rock bolts in engineering.

Classification of anchor bolts based on spectral kurtosis and K-means clustering algorithm

In nondestructive testing of anchor bolt quality, it is important to identify the state of the anchor bolt accurately. In this paper, a combination method of spectral kurtosis and K-means clustering algorithm is proposed to identify different types of anchor models: maximum and nonzero minimum values are extracted as eigenvalues from spectral kurtosis distribution of anchor model signals which was calculated through fast kurtogram algorithm, then the eigenvalues are automatically classified by K-means clustering algorithm to realize the identification of different anchor bolt model. Through verification test, this method is proved to be fast, effective and high accuracy for anchor bolts classification.

Study of rock bolts detection based on fusion algorithm

Length is an important factor to evaluate the quality of rock bolts. Due to the harsh environment and large amounts of noise contained by rock signals detection, its difficult to estimate the rock bolts parameters. In this paper, a non-destructive testing(NDT) method is used to detect the parameters of free bolt. Six different types of excitation signal is used on rock bolts, the reflection signal is detected respectively by ultrasonic sensor, which is analysed by a fusion algorithm of D-S evidence theory, the result shows that this method is satisfied.

The study of nondestructive testing of rock bolts based on PNN and wavelet packet

Anchoring technology is widely used in slope, tunnels and underground engineering. However, the quality of rock bolts is still a hot problem difficult to solve. Considering the shortcoming of pull-out testing, defect recognition in a nondestructive way is necessary. Decomposing the signals obtained by bolt quality detector with wavelet packet; extracting energy feature by wavelet packet energy spectrum; converting the normalized energy eigenvector as input of probabilistic neural network. With a higher accuracy than RBF, the PNN model can provide a reference for recognition defects of rock bolts in engineering without destruction.

Online fault detection of networked control systems

A method for online fault detection of networked control system considering the network-induced delay was studied in this paper. First of all, a fault detection model was established for the impact of the unknown network-induced delay. Based on the model, the fault detection filter was constructed, and then the fault detection is converted to a optimization problem. The numerical simulation shows that the proposed online fault detection algorithm is not only sensitive to the fault, but also robust to the unknown disturbance caused by the delay.