Science and Engineering is an international, peer-reviewed open access journal dedicated to advancing research the field of science and engineering. The journal provides a rapid publication process to ensure wide dissemination of high-quality articles to scientists, professionals, and interested individuals worldwide. Our goal is to serve as an efficient, reliable, and trusted platform for scholars and readers, publishing cutting-edge research in the field.
Abstract: In this paper, we study the fault diagnosis problem for a class of discrete nonlinear systems with unmeasurable premise variables, which are Takagi-Sugeno (T-S) models. The studied T-S structure can simplify the computation by using rules containing local nonlinearities to reduce the number of rules in the model. In designing the fault detection observer, the main consideration is the unknown input signal, while maximizing the effect of actuator failure on the generated residuals in order to minimize the impact of uncertainty on the system performance. The H- performance index of residual-to-fault sensitivity and the H∞ performance index of residual-to-unknown-input robustness are described, and the design problem of a robust fault detection observer is described as an optimal design problem satisfying H∞/H-. The unmeasurable premise variables are compared to the case of measurable premise variables, which can represent larger nonlinear systems. Then, a new iterative linear matrix inequality (LMI) algorithm and convex optimization technique are proposed to solve the optimal observer gain matrix by giving the existence conditions of the fault detection observer. Finally, the effectiveness of the method is verified by numerical examples, and the fault detection observer designed by this method is highly sensitive to faults and robust to unknown inputs.Abstract: In this paper, we study the fault diagnosis problem for a class of discrete nonlinear systems with unmeasurable premise variables, which are Takagi-Sugeno (T-S) models. The studied T-S structure can simplify the computation by using rules containing local nonlinearities to reduce the number of rules in the model. In designing the fault detection ob...Learn More
Abstract: The J-A two-parameter method proposed for the asymptotic analytical solution of type-I crack-tip field can effectively characterize the crack-tip stress field distribution and its constraint for power-hardening materials. This paper systematically summarizes commonly used methods for characterizing stress fields at crack tips. Taking the single-edged notched tension specimen as an example, the superiority of the J-A two-parameter method in characterizing the crack tip constraint of type I cracks is demonstrated. The J-A two-parameter method consists of three important parameters, including characteristic index, dimensionless integration constant and stress distribution angle function. Based on the existing high-order analytical solutions for the crack tip asymptotic field of power-hardening materials, a calculation program is compiled to compute these important parameters. High-precision empirical formulas for various parameters, ranging from low to high hardening levels, are obtained using the nonlinear fitting method with the determination coefficient R2 reaching above 0.999. Finally, two methods for calculating the constraint parameter A are presented. It is found that using the fitting method to determine the constraint parameter A in the fracture control zone is more reasonable. A formula for directly calculating parameter A is also provided. Combined with the fitting formulas for the three parameters, the crack tip constraint of the single-edged notched tension specimen is calculated. Results show that the proposed calculation formulas can obtain crack tip constraint parameter A with high precision, facilitating the wide application of the J-A two-parameter method.Abstract: The J-A two-parameter method proposed for the asymptotic analytical solution of type-I crack-tip field can effectively characterize the crack-tip stress field distribution and its constraint for power-hardening materials. This paper systematically summarizes commonly used methods for characterizing stress fields at crack tips. Taking the single-edg...Learn More
Abstract: The compressive strength of concrete is one of the main factors affecting the quality of construction, and it is also a key parameter for concrete design, so it is necessary to predict it. However, most of the early empirical formulas for predicting the compressive strength of concrete are for specific concretes, which can not be universally promoted, while other prediction methods such as physical testing method, finite element analysis method, etc. are cumbersome in terms of steps, and can not be formed into a generalized prediction model under the harsh conditions of applicability. In recent years, the concrete compressive strength prediction model based on machine learning has become a hot research topic at home and abroad. In this paper, firstly, by comparing the application of various machine learning methods, including regression model, support vector regression machine model, tree model, neural network and integrated learning in concrete compressive strength, it is concluded that the support vector regression machine model and integrated learning have higher effectiveness in the prediction of compressive strength. Finally, the article discusses the shortcomings of the widely used concrete compressive strength prediction methods and gives the future research direction of machine learning based concrete compressive strength prediction model.Abstract: The compressive strength of concrete is one of the main factors affecting the quality of construction, and it is also a key parameter for concrete design, so it is necessary to predict it. However, most of the early empirical formulas for predicting the compressive strength of concrete are for specific concretes, which can not be universally promot...Learn More
