Performance Of Diagnosis Model Research Articles

A Cost-Sensitive Diagnosis Method Based on the Operation and Maintenance Data of UAV

In the fault diagnosis of UAVs, extremely imbalanced data distribution and vast differences in effects of fault modes can drastically affect the application effect of a data-driven fault diagnosis model under the limitation of computing resources. At present, there is still no credible approach to determine the cost of the misdiagnosis of different fault modes that accounts for the interference of data distribution. The performance of the original cost-insensitive flight data-driven fault diagnosis models also needs to be improved. In response to this requirement, this paper proposes a two-step ensemble cost-sensitive diagnosis method based on the operation and maintenance data of UAV. According to the fault criticality from FMECA information, we defined a misdiagnosis hazard value and calculated the misdiagnosis cost. By using the misdiagnosis cost, a static cost matrix could be set to modify the diagnosis model and to evaluate the performance of the diagnosis results. A two-step ensemble cost-sensitive method based on the MetaCost framework was proposed using stratified bootstrapping, choosing LightGBM as meta-classifiers, and adjusting the ensemble form to enhance the overall performance of the diagnosis model and reduce the occupation of the computing resources while optimizing the total misdiagnosis cost. The experimental results based on the KPG component data of a large fixed-wing UAV show that the proposed cost-sensitive model can effectively reduce the total cost incurred by misdiagnosis, without putting forward excessive requirements on the computing equipment under the condition of ensuring a certain overall level of diagnosis performance.

A Novel Self-Decision Fault Diagnosis Model Based on State-Oriented Correction for Power Transformer

Dissolved gas analysis can provide powerful technical support for assessment and fault diagnosis of power transformers. However, due to inadaptability of model structure, the performance of diagnosis model can hardly be improved with theoretical support. To fill the gap, a novel self-decision fault diagnosis model for power transformer has been proposed, taking into consideration of both the characteristics of faults and adaptability of conventional deep brief network. Specifically, an active-correction unit, as the core of this novel model, was established to detect the real-time state of the diagnosis model and then select the corresponding correction strategy. An error neuron is also placed to serve as the starting point of the state-oriented correction and load the correction signal that is sent from the active-correction unit. Verifications in the field indicates that the fault correlation extracted by the sparse autoencoder can eliminate error in the training process. When the results of IEC three-ratios method is integrated into the input neuron, the accuracy of proposed model can reach more than 92.3%. Based on the proven effectiveness and feasibility of the proposed method in state-oriented correction, this method can provide a reliable support for substation preventive maintenance.

Rolling Bearing Fault Diagnosis Based on Sensitive Feature Transfer Learning and Local Maximum Margin Criterion under Variable Working Condition

In real industrial scenarios, the working conditions of bearings are variable, and it is therefore difficult for data-driven diagnosis methods based on conventional machine-learning techniques to guarantee the desirable performance of diagnosis models, as the models assume that the distributions of both the training and testing data are the same. To enhance the performance of the fault diagnosis of bearings under different working conditions, a novel diagnosis framework inspired by feature extraction, transfer learning (TL), and feature dimensionality reduction is proposed in this work, and dual-tree complex wavelet packet transform (DTCWPT) is used for signal processing. Additionally, transferable sensitive feature selection by ReliefF and the sum of mean deviation (TSFSR) is proposed to reduce the redundant information of the original feature set, to select sensitive features for fault diagnosis, and to reduce the difference between the marginal distributions of the training and testing feature sets. Furthermore, a modified feature reduction method, the local maximum margin criterion (LMMC), is proposed to acquire low-dimensional mapping for high-dimensional feature spaces. Finally, bearing vibration signals collected from two test rigs are analyzed to demonstrate the adaptability, effectiveness, and practicability of the proposed diagnosis framework. The experimental results show that the proposed method can achieve high diagnosis accuracy and has significant potential benefits in industrial applications.

A New DGA Based Transformer Fault Diagnosis Scheme Suitable for Time-Series Fault Data

The quality of original data is crucial to the performance of diagnosis model. To improve the performance of transformer diagnosis model based on Dissolved Gas Analysis (DGA), a new diagnosis scheme suitable for time-series dissolved gas data is proposed in this paper. After the analysis of traditional transformer diagnosis architecture, a fault data extraction step is added to the architecture to improve the quality of original fault data. The fault data extraction step is mainly composed of two parts, invalid data correction and determination of possible initial fault time based on fault early warning. Finally, the numerical results validate that the accuracy and sensitivity of DGA based fault diagnosis for the transformer are improved by extracting fault feature of time-series data.

Use of a neural network to link AFM etch patterns to plasma parameters

Plasmas play a critical role in depositing thin films or etching fine patterns while manufacturing integrated circuits. A new model for plasma diagnosis is presented. This was accomplished by linking atomic force microscopy (AFM) to plasma parameters using a neural network. Experimental AFM data were collected during the etching of silicon oxynitride films in C 2F 6 inductively coupled plasma. Surface roughness of etched patterns was characterized by means of discrete wavelet transformation. This led to the construction of three vertical (type I), diagonal (type II), and horizontal (type III) wavelet coefficient-based models. The performance of diagnosis models was evaluated in terms of the prediction and recognition accuracies. Both accuracies were optimized as a function of the number of hidden neurons. Comparisons revealed that the type I model yielded the largest recognition and the smallest prediction error. This was demonstrated even under stricter monitoring conditions. More improved diagnosis is expected by enhancing AFM resolution.