Search results

The purpose of this study is to simultaneously determine the constitutive parameters and boundary conditions by solving inverse mechanical problems of power hardening elastoplastic materials in three-dimensional geometries.

The power hardening elastoplastic problem is solved by the complex variable finite element method in software ABAQUS, based on a three-dimensional complex stress element using user-defined element subroutine. The complex-variable-differentiation method is introduced and used to accurately calculate the sensitivity coefficients in the multiple parameters identification method, and the Levenberg–Marquardt algorithm is applied to carry out the inversion.

Numerical results indicate that the complex variable finite element method has good performance for solving elastoplastic problems of three-dimensional geometries. The inversion method is effective and accurate for simultaneously identifying multi-parameters of power hardening elastoplastic problems in three-dimensional geometries, which could be employed for solving inverse elastoplastic problems in engineering applications.

The constitutive parameters and boundary conditions are simultaneously identified for power hardening elastoplastic problems in three-dimensional geometries, which is much challenging in practical applications. The numerical results show that the inversion method has high accuracy, good stability, and fast convergence speed.

This paper aims to present the Cauchy problem for the Laplace’s equation for profiles of gas turbine blades with one and three cooling channels. The distribution of heat transfer coefficient and temperature on the outer boundary of the blade are known. On this basis, the temperature on inner surfaces of the blade (the walls of cooling channels) is determined.

Such posed inverse problem was solved using the finite element method in the domain of the discrete Fourier transform (DFT).

Calculations indicate that the regularization in the domain of the DFT enables obtaining a stable solution to the inverse problem. In the example under consideration, problems with reconstruction constant temperature, assumed on the outer boundary of the blade, in the vicinity of the trailing and leading edges occurred.

The application of DFT in connection with regularization is an original achievement presented in this study.

Inverse problems in electromagnetism, namely, the recovery of sources (currents or charges) or system data from measured effects, are usually ill-posed or, in the numerical formulation, ill-conditioned and require suitable regularization to provide meaningful results. To test new regularization methods, there is the need of benchmark problems, which numerical properties and solutions should be well known. Hence, this study aims to define a benchmark problem, suitable to test new regularization approaches and solves with different methods.

To assess reliability and performance of different solving strategies for inverse source problems, a benchmark problem of current synthesis is defined and solved by means of several regularization methods in a comparative way; subsequently, an approach in terms of an artificial neural network (ANN) is considered as a viable alternative to classical regularization schemes. The solution of the underlying forward problem is based on a finite element analysis.

The paper provides a very detailed analysis of the proposed inverse problem in terms of numerical properties of the lead field matrix. The solutions found by different regularization approaches and an ANN method are provided, showing the performance of the applied methods and the numerical issues of the benchmark problem.

The value of the paper is to provide the numerical characteristics and issues of the proposed benchmark problem in a comprehensive way, by means of a wide variety of regularization methods and an ANN approach.

The purpose of this paper is to develop a transform method and a deep learning model to identify the inner surface shape based on the measurement temperature at the outer boundary of the pipe.

The training process is assisted by the finite element method (FEM) simulation which solves the direct problem for the data preparation. To avoid re-meshing the domain when the inner surface shape varies, a new transform method is proposed to transform the shape identification problem into the effective thermal conductivity identification problem. The deep learning model is established to set up the relationship between the measurement temperature and the effective thermal conductivity. Then the unknown geometry shape is acquired by the mapping between the inner shape and the effective thermal conductivity through the inverse transform method.

The new method is successfully applied to identify the internal boundary of a pipe with eccentric circle, ellipse and nephroid inner geometries. The results show that as the measurement points increased and the measurement error decreased, the results became more accurate. The position of the measurement point and mesh density of the FEM model have less effect on the results.

The deep learning model and the transform method are developed to identify the pipe inner surface shape. There is no need to re-mesh the domain during the computation progress. The results show that the proposed method is a fast and an accurate tool for identifying the pipe inner surface.

Many a times, the information about the boundary heat flux is obtained only through inverse approach by locating the thermocouple or temperature sensor in accessible boundary. Most of the work reported in literature for the estimation of unknown parameters is based on heat conduction model. Inverse approach using conjugate heat transfer is found inadequate in literature. Therefore, the purpose of the paper is to develop a 3D conjugate heat transfer model without model reduction for the estimation of heat flux and heat transfer coefficient from the measured temperatures.

A 3 D conjugate fin heat transfer model is solved using commercial software for the known boundary conditions. Navier–Stokes equation is solved to obtain the necessary temperature distribution of the fin. Later, the complete model is replaced with neural network to expedite the computations of the forward problem. For the inverse approach, genetic algorithm (GA) and particle swarm optimization (PSO) are applied to estimate the unknown parameters. Eventually, a hybrid algorithm is proposed by combining PSO with Broyden–Fletcher–Goldfarb–Shanno (BFGS) method that outperforms GA and PSO.

The authors demonstrate that the evolutionary algorithms can be used to obtain accurate results from simulated measurements. Efficacy of the hybrid algorithm is established using real time measurements. The hybrid algorithm (PSO-BFGS) is more efficient in the estimation of unknown parameters for experimentally measured temperature data compared to GA and PSO algorithms.

Surrogate model using ANN based on computational fluid dynamics simulations and in-house steady state fin experiments to estimate the heat flux and heat transfer coefficient separately using GA, PSO and PSO-BFGS.

The purpose of this paper is to apply the meshless method of fundamental solutions (MFS) to the two‐dimensional time‐dependent heat equation in order to locate an unknown internal inclusion.

The problem is formulated as an inverse geometric problem, using non‐invasive Dirichlet and Neumann exterior boundary data to find the internal boundary using a non‐linear least‐squares minimisation approach. The solver will be tested when locating a variety of internal formations.

The method implemented was proven to be both stable and reasonably accurate when data were contaminated with random noise.

Owing to limited computational time, spatial resolution of internal boundaries may be lower than some similar case investigations.

This research will have practical implications to the modelling and monitoring of crystalline deposit formations within the nuclear industry, allowing development of future designs.

Similar work has been completed in regards to the steady state heat equation, however to the best of the authors' knowledge no previous work has been completed on a time‐dependent inverse inclusion problem relating to the heat equation, using the MFS. Preliminary results presented here will have value for possible future design and monitoring within the nuclear industry

Prediction of excess pore water pressure and estimation of soil parameters are the two key interests for consolidation problems, which can be mathematically quantified by a set of partial differential equations (PDEs). Generally, there are challenges in solving these two issues using traditional numerical algorithms, while the conventional data-driven methods require massive data sets for training and exhibit negative generalization potential. This paper aims to employ the physics-informed neural networks (PINNs) for solving both the forward and inverse problems.

A typical consolidation problem with continuous drainage boundary conditions is firstly considered. The PINNs, analytical, and finite difference method (FDM) solutions are compared for the forward problem, and the estimation of the interface parameters involved in the problem is discussed for the inverse problem. Furthermore, the authors also explore the effects of hyperparameters and noisy data on the performance of forward and inverse problems, respectively. Finally, the PINNs method is applied to the more complex consolidation problems.

The overall results indicate the excellent performance of the PINNs method in solving consolidation problems with various drainage conditions. The PINNs can provide new ideas with a broad application prospect to solve PDEs in the field of geotechnical engineering, and also exhibit a certain degree of noise resistance for estimating the soil parameters.

This study presents the potential application of PINNs for the consolidation of soils. Such a machine learning algorithm helps to obtain remarkably accurate solutions and reliable parameter estimations with fewer and average-quality data, which is beneficial in engineering practice.

For efficient trajectory control of industrial robots, a cumbersome computation for inverse kinematics and inverse dynamics is needed, which is usually developed using spatial transformation using Denavit–Hartenberg principle and Lagrangian or Newton–Euler methods, respectively. The model is highly non-linear and needs to deal with uncertainties because of lack of accurate measurement of mechanical parameters, noise and non-inclusion of joint friction, which results in some inaccuracies in predicting accurate torque trajectories. To get a guaranteed closed form solution, the robot designers normally follow Pieper’s recommendation and compromise with the mechanical design. While this may be acceptable for the industrial robots where the aesthetic look is not that important, it is not for humanoid and social robots. To help solve this problem, this study aims to propose an alternative machine learning-based computational approach based on a multi-gated sequence model for finding appropriate mapping between Cartesian space to joint space and motion space to joint torque space.

First, the authors generate sufficient data required for the sequence model, using forward kinematics and forward dynamics by running N number of nested loops, where N is the number of joints of the robot. Subsequently, to develop a learning-based model based on sequence analysis, the authors propose to use long short-term memory (LSTM) and hence, train an LSTM model, the architecture details of which have been discussed in the paper. To make LSTM learning algorithms perform efficiently, the authors need to detect and eliminate redundant features from the data set, which the authors propose to do using an elegant statistical tool called Pearson coefficient.

To validate the proposed model, the authors have performed rigorous experiments using both hardware and simulation robots (Baxter/Anukul robot) available in their laboratory and KUKA simulation robot data set made available from Neural Learning for Robotics Laboratory. Through several characteristic plots, it has been shown that a sequence-based LSTM model of deep learning architecture with non-redundant features could help the robots to learn smooth and accurate trajectories more quickly compared to data sets having redundancy. Such data-driven modeling techniques can change the future course of direction of robotics research for solving the classical problems such as trajectory planning and motion planning for manipulating industrial as well as social humanoid robots.

The present investigation involves development of deep learning-based computation model, statistical analyses to eliminate redundant features, data creation from one hardware robot (Anukul) and one simulation robot model (KUKA), rigorously training and testing separately two computational models (specially configured two LSTM models) – one for learning inverse kinematics and one for learning inverse dynamics problem – and comparison of the inverse dynamics model with the state-of-the-art model. Hence, the authors strongly believe that the present paper is compact and complete to get published in a reputed journal so that dissemination of new ideas can benefit the researchers in the area of robotics.

Draws a comparison between the use of a genetic algorithm and the sequential function specification method for the solution of a one‐dimensional linear inverse thermal field problem, based on the use of noisy measurements. In solving this problem aims to estimate the value of a single constant convective heat transfer coefficient. Documents the findings that both approaches can provide estimates within 1 per cent of the target solution and that the sensitivity and robustness of each approach to measurement location, time step size and measurement errors are markedly different.