Angular Acceleration Sensor Fault Diagnosis Based on LM-BP Neural Network

Liu, Hua; Li, Bo; Liu, Tong; Wang, Meiling; Fu, Huijin; Guo, Ruoyu

doi:10.23919/chicc.2018.8484216

Cited by 3 publications

(2 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The method has quadratic convergence, like Newton's method, if the Jacobi matrix is nondegenerate and Lipschitz continuous. It has become a standard method not only for solving nonlinear least squares problems, but is also widely used in solving other problems, such as in artificial neural networks [28], robotics [29], speech recognition systems [30], optimization of technological processes [31], and algorithms for determining the coordinates and orientation of an object [32]. The method has a disadvantage in that at each step of the iterative process it is necessary, in addition to the gradient of the function, to calculate the Hessian and solve the problem of matrix inversion.…”

Section: Introductionmentioning

confidence: 99%

Modeling the Kinetics of the Singlet Oxygen Effect in Aqueous Solutions of Proteins Exposed to Thermal and Laser Radiation

et al. 2022

View full text Add to dashboard Cite

A system of kinetic equations describing the changes in the concentration of reactive oxygen species (ROS) in aqueous solutions of proteins was obtained from the analysis of chemical reactions involving singlet oxygen. Applying the condition of the stationarity of the intermediate products to the system, we determined the functional dependence of the hydrogen peroxide concentration on the protein concentration under the action of thermal and laser radiation. An approximate analytical solution to the nonlinear system of differential equations that define the ROS concentration dynamics was found. For aqueous solutions of bovine serum albumin (BSA) and bovine gamma globulin (BGG), the orders and rate constants of the reactions describing the ROS conversions were determined by minimizing the sum of squared deviations of the functions found by solving both the static and dynamic problems from experimentally measured dependences. When solving the optimization problem, the Levenberg–Marquardt algorithm was used.

show abstract

Section: Introductionmentioning

confidence: 99%

Modeling the Kinetics of the Singlet Oxygen Effect in Aqueous Solutions of Proteins Exposed to Thermal and Laser Radiation

et al. 2022

View full text Add to dashboard Cite

show abstract

“…In Refs. [9,10], the wavelet method was used to preprocess data and the LM neural networks method was used to analyze them. In Ref.…”

Section: Introductionmentioning

confidence: 99%

Adaptive Levenberg–Marquardt Algorithm: A New Optimization Strategy for Levenberg–Marquardt Neural Networks

et al. 2021

View full text Add to dashboard Cite

Engineering data are often highly nonlinear and contain high-frequency noise, so the Levenberg–Marquardt (LM) algorithm may not converge when a neural network optimized by the algorithm is trained with engineering data. In this work, we analyzed the reasons for the LM neural network’s poor convergence commonly associated with the LM algorithm. Specifically, the effects of different activation functions such as Sigmoid, Tanh, Rectified Linear Unit (RELU) and Parametric Rectified Linear Unit (PRLU) were evaluated on the general performance of LM neural networks, and special values of LM neural network parameters were found that could make the LM algorithm converge poorly. We proposed an adaptive LM (AdaLM) algorithm to solve the problem of the LM algorithm. The algorithm coordinates the descent direction and the descent step by the iteration number, which can prevent falling into the local minimum value and avoid the influence of the parameter state of LM neural networks. We compared the AdaLM algorithm with the traditional LM algorithm and its variants in terms of accuracy and speed in the context of testing common datasets and aero-engine data, and the results verified the effectiveness of the AdaLM algorithm.

show abstract

Gas sensor fault diagnosis for imbalanced data based on generative adversarial networks

Sun

Liu

et al. 2021

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

Gas sensors play a key role in gas detection. Gas sensors are prone to failure, so they are of great significance to classify gas sensor fault. Traditional gas sensor fault diagnosis methods do not consider the condition of the imbalance of fault data. In this paper, the generative adversarial networks (GAN) method is used to generate small sample sensor fault data, balance various types of sensor fault data, and then the random forest is used for fault classification. Finally, the sensor fault data is obtained by self-made experimental system to verify the proposed method. The experimental results show that the proposed method could effectively improve the accuracy of gas sensor fault diagnosis.

show abstract

Angular Acceleration Sensor Fault Diagnosis Based on LM-BP Neural Network

Cited by 3 publications

References 5 publications

Modeling the Kinetics of the Singlet Oxygen Effect in Aqueous Solutions of Proteins Exposed to Thermal and Laser Radiation

Modeling the Kinetics of the Singlet Oxygen Effect in Aqueous Solutions of Proteins Exposed to Thermal and Laser Radiation

Adaptive Levenberg–Marquardt Algorithm: A New Optimization Strategy for Levenberg–Marquardt Neural Networks

Gas sensor fault diagnosis for imbalanced data based on generative adversarial networks

Contact Info

Product

Resources

About