A New Non-Parametric Model Based on Neural Network for a MR Damper

Currently dampers based on magnetorheological (MR) fluids are being used in many applications such as construction, biomechanical and semi-active suspension to improve their behaviour. The main advantage of MR dampers is its very low time response (≈ 10 ms). In many cases, it is necessary to establish a suitable model of MR damper which characterizes its behaviour so that this model can be used in the simulation stage. In this paper, a new non-parametric model is proposed based on neural networks using a recursive lazy learning to model the MR damper behaviour. The proposed method is validated by comparison with experimental obtained responses. Results show that the estimated model correlates very well with the data obtained experimentally and learns quickly.

Download Full-text

A Non Parametric Model for Magneto Rheological Dampers

Volume 7B: 17th Biennial Conference on Mechanical Vibration and Noise ◽

10.1115/detc99/vib-8017 ◽

1999 ◽

Author(s):

Mehdi Ahmadian ◽

Xubin Song

Keyword(s):

Test Data ◽

Mr Damper ◽

Parametric Model ◽

Parametric Models ◽

Magnetic Saturation ◽

Mr Dampers ◽

Electro Magnetic ◽

Magneto Rheological ◽

Force Characteristics ◽

Non Parametric

Abstract A non-parametric model for magneto-rheological (MR) dampers is presented. After discussing the merits of parametric and non-parametric models for MR dampers, the test data for a MR damper is used to develop a non-parametric model. The results of the model are compared with the test data to illustrate the accuracy of the model. The comparison shows that the non-parametric model is able to accurately predict the damper force characteristics, including the damper non-linearity and electro-magnetic saturation. It is further shown that the parametric model can be numerically solved more efficiently than the parametric models.

Download Full-text

An investigation into the use of neural networks for the semi-active control of a magnetorheologically damped vehicle suspension

Proceedings of the Institution of Mechanical Engineers Part D Journal of Automobile Engineering ◽

10.1243/09544070jauto1481 ◽

2010 ◽

Vol 224 (7) ◽

pp. 829-848 ◽

Cited By ~ 24

Author(s):

H Metered ◽

P Bonello ◽

S O Oyadiji

Keyword(s):

Neural Network ◽

Neural Networks ◽

Inverse Dynamics ◽

Ride Comfort ◽

Mr Damper ◽

Performance Criteria ◽

Vehicle Suspension ◽

Neural Network Controller ◽

Road Disturbance ◽

Continuous State

Neural networks are highly useful for the modelling and control of magnetorheological (MR) dampers. A damper controller based on a recurrent neural network (RNN) of the inverse dynamics of an MR damper potentially offers significant advantages over conventional controllers in terms of reliability and cost through the minimal use of sensors. This paper introduces a neural-network-based MR damper controller for use in conjunction with the system controller of a semi-active vehicle suspension. A mathematical model of a semi-active quarter-vehicle suspension using an MR damper is derived. Control performance criteria are evaluated in the time and frequency domains in order to quantify the suspension effectiveness under bump and random road disturbance. Studies using the modified Bouc—Wen model for the MR damper as well as an actual damper fitted in a hardware-in-the-loop simulation (HILS) both showed that the inverse RNN damper controller potentially offers a significantly superior ride comfort and vehicle stability over a conventional MR damper controller based on continuous-state control. The neural network controller produces a smoother and lower input voltage to the MR damper coil, ensuring extended damper life and lower power requirement respectively. Further studies performed using an RNN model of the forward dynamics of the MR damper showed that it is a reliable substitute for HILS for validating multi-damper control applications.

Download Full-text

Flexible, non-parametric modeling using regularized neural networks

Computational Statistics ◽

10.1007/s00180-021-01190-4 ◽

2022 ◽

Author(s):

Oskar Allerbo ◽

Rebecka Jörnsten

Keyword(s):

Neural Network ◽

Neural Networks ◽

Heterogeneous Data ◽

Parametric Modeling ◽

Additive Models ◽

Adaptive Lasso ◽

Complex Data ◽

Data Set ◽

Model Complex ◽

Non Parametric

AbstractNon-parametric, additive models are able to capture complex data dependencies in a flexible, yet interpretable way. However, choosing the format of the additive components often requires non-trivial data exploration. Here, as an alternative, we propose PrAda-net, a one-hidden-layer neural network, trained with proximal gradient descent and adaptive lasso. PrAda-net automatically adjusts the size and architecture of the neural network to reflect the complexity and structure of the data. The compact network obtained by PrAda-net can be translated to additive model components, making it suitable for non-parametric statistical modelling with automatic model selection. We demonstrate PrAda-net on simulated data, where we compare the test error performance, variable importance and variable subset identification properties of PrAda-net to other lasso-based regularization approaches for neural networks. We also apply PrAda-net to the massive U.K. black smoke data set, to demonstrate how PrAda-net can be used to model complex and heterogeneous data with spatial and temporal components. In contrast to classical, statistical non-parametric approaches, PrAda-net requires no preliminary modeling to select the functional forms of the additive components, yet still results in an interpretable model representation.

Download Full-text

Two-Dimensional CFD Modeling of Hysteresis Behavior of MR Dampers

Shock and Vibration ◽

10.1155/2019/9383047 ◽

2019 ◽

Vol 2019 ◽

pp. 1-14 ◽

Cited By ~ 4

Author(s):

Pengfei Guo ◽

Jing Xie

Keyword(s):

Mr Damper ◽

Cfd Modeling ◽

Two Dimensional ◽

The Finite Element Method ◽

One Dimensional ◽

Hysteresis Behavior ◽

Mr Dampers ◽

Nonlinear Hysteresis ◽

Mr Fluids ◽

Dimensional Modeling

So far, most previous studies on the nonlinear hysteresis analysis of ER/MR dampers have been limited to one-dimensional modeling techniques. A two-dimensional (2D) axisymmetric CFD model of MR dampers is developed in this work. The main advantage of the proposed 2D model of MR dampers lies in that it can be used to predict dynamic behavior of MR devices of arbitrary geometries. The compressibility of MR fluids is the main factor responsible for the hysteresis behavior of MR dampers, and it has been rarely investigated in previous multidimensional modeling of MR damper. In our model, the compressibility of MR fluids is taken into account by the two-dimensional constitutive model which is extended from a previous one-dimensional physical model. The model is solved by using the finite element method, and the movement of the piston is described by the moving mesh technique. The MR damper in a reference is simulated, and the model predictions show good agreement with the experimental data in the reference.

Download Full-text

Non-Parametric Model of Magneto-Rheological (MR) Damper Based on Adaptive Neuro-Fuzzy Inference System (ANFIS)

Advanced Engineering Forum ◽

10.4028/www.scientific.net/aef.2-3.1059 ◽

2011 ◽

Vol 2-3 ◽

pp. 1059-1066 ◽

Cited By ~ 1

Author(s):

Ling Zheng ◽

Fei Liu ◽

Zhong Yong Zhou

Keyword(s):

Fuzzy Inference ◽

Mr Damper ◽

Parametric Model ◽

Control Voltage ◽

Nonlinear Relationship ◽

Damping Force ◽

Inference System ◽

Neuro Fuzzy ◽

Magneto Rheological ◽

Non Parametric

Magneto-rheological (MR) damper is a typical non-linear actuator. The nonlinear relationship between the input and the output of MR damper should be characterized accurately in order to produce a required control action supplied by MR damper. The challenges of the conventional parametric model are the time-consuming and the requirement of complex parameter identification for MR damper. In this paper, a non-parametric model of MR damper based on the adaptive neuro-fuzzy system theory is presented to overcome the drawbacks of the conventional parametric model. This non-parametric model is developed by means of two adaptive neural fuzzy sub-systems which are designed to describe the nonlinear relationship in MR damper successfully. One sub-system is used to characterize the dependence of the damping force on velocity and acceleration, the other sub-system is used to characterize the dependence of the damping force level on the control voltage. The proposed non-parametric model is identified by experimental results. Furthermore, the accuracy of the model is investigated and evaluated. The model supplies a key technical support to achieve excellent control performances for semi-active suspension systems with MR dampers in vehicle.

Download Full-text

Neural Network – Backstepping Control for Vibration Reduction in a Magnetorheological Suspension System

Solid State Phenomena ◽

10.4028/www.scientific.net/ssp.147-149.839 ◽

2009 ◽

Vol 147-149 ◽

pp. 839-844 ◽

Cited By ~ 4

Author(s):

Mauricio Zapateiro ◽

Ning Su Luo ◽

Hamid Reza Harimi

Keyword(s):

Neural Network ◽

Inverse Dynamics ◽

Control Variable ◽

Mr Damper ◽

Suspension System ◽

Control Voltage ◽

Damping Force ◽

Mr Dampers ◽

Vehicle Suspension System ◽

Nonlinear Devices

In this paper we address the design of the controller for semi-active vehicle suspension system that employs an MR damper as the actuator. MR dampers are nonlinear devices which are difficult to model. Several MR damper forward models have been proposed; they can estimate the damping force of the device taking variables such as control voltage and velocity inputs. However, the inverse model, i.e., the model that computes the control variable is even more difficult to find due to the numerical complexity that implies the inverse of the nonlinear forward model. In our case, we develop a neural network able to reproduce such inverse dynamics. This neural network is connected to a backstepping controller that estimates the damping force to reduce the vibrations of the system. The performance of the controller is evaluated by means of simulations in MATLAB/Simulink.

Download Full-text

Optimal design of magnetorheological damper with multiple axial fluid flow channels using BP neural network and particle swarm optimization methodologies

International Journal of Applied Electromagnetics and Mechanics ◽

10.3233/jae-210089 ◽

2021 ◽

pp. 1-22

Author(s):

Guoliang Hu ◽

Haonan Qi ◽

Miao Chen ◽

Lifan Yu ◽

Gang Li ◽

...

Keyword(s):

Neural Network ◽

Particle Swarm Optimization ◽

Fluid Flow ◽

Magnetic Flux ◽

Bp Neural Network ◽

Particle Swarm ◽

Mr Damper ◽

Swarm Optimization ◽

Mr Dampers ◽

Flow Channels

In this paper, a magnetorheological (MR) damper with multiple axial fluid flow channels is developed to solve the conflicts between limitation of size dimension and improvement of damping performance. By setting symmetrical excitation coils at both ends of the MR damper, the effective fluid flow channels of the proposed MR damper are significantly lengthened. In order to investigate the distributions of magnetic flux lines and magnetic flux density of the MR damper, the finite element model of the MR damper is established by using ANSYS software. Moreover, an optimization method combining BP neural network and particle swarm optimization (PSO) is proposed to improve the magnetic field utilization of the designed damper, and the damping performances of initial and optimal MR dampers are also experimentally tested. The test results show that the output damping force of initial and optimal MR dampers is 3.13 kN and 5.98 kN respectively under the applied current of 1.8 A, increasing by 91.1%, and the dynamic adjustable range is 11.5 and 16.1 respectively, increasing by 40.0%. It is found that the damping performance of the proposed MR damper is significantly improved.

Download Full-text

USING MULTIVARIATE ADAPTIVE REGRESSION SPLINE AND ARTIFICIAL NEURAL NETWORK TO SIMULATE URBANIZATION IN MUMBAI, INDIA

ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences ◽

10.5194/isprsarchives-xl-1-w5-31-2015 ◽

2015 ◽

Vol XL-1-W5 ◽

pp. 31-36 ◽

Cited By ~ 6

Author(s):

M. Ahmadlou ◽

M. R. Delavar ◽

A. Tayyebi ◽

H. Shafizadeh-Moghadam

Keyword(s):

Neural Network ◽

Artificial Neural Network ◽

Urban Areas ◽

Parametric Model ◽

Parametric Models ◽

Model Structure ◽

Regression Spline ◽

Adaptive Regression ◽

Modelling Process ◽

Non Parametric

Land use change (LUC) models used for modelling urban growth are different in structure and performance. Local models divide the data into separate subsets and fit distinct models on each of the subsets. Non-parametric models are data driven and usually do not have a fixed model structure or model structure is unknown before the modelling process. On the other hand, global models perform modelling using all the available data. In addition, parametric models have a fixed structure before the modelling process and they are model driven. Since few studies have compared local non-parametric models with global parametric models, this study compares a local non-parametric model called multivariate adaptive regression spline (MARS), and a global parametric model called artificial neural network (ANN) to simulate urbanization in Mumbai, India. Both models determine the relationship between a dependent variable and multiple independent variables. We used receiver operating characteristic (ROC) to compare the power of the both models for simulating urbanization. Landsat images of 1991 (TM) and 2010 (ETM<sup>+</sup>) were used for modelling the urbanization process. The drivers considered for urbanization in this area were distance to urban areas, urban density, distance to roads, distance to water, distance to forest, distance to railway, distance to central business district, number of agricultural cells in a 7 by 7 neighbourhoods, and slope in 1991. The results showed that the area under the ROC curve for MARS and ANN was 94.77% and 95.36%, respectively. Thus, ANN performed slightly better than MARS to simulate urban areas in Mumbai, India.

Download Full-text

SCORING MODELING BASED ON NEURAL NETWORKS FOR DETERMINING A BANK BORROWER'S RATING

Economy of Ukraine ◽

10.15407/economyukr.2020.10.054 ◽

2020 ◽

Vol 2020 (10) ◽

pp. 54-62

Author(s):

Oleksii VASYLIEV ◽

Keyword(s):

Neural Network ◽

Neural Networks ◽

Network Architecture ◽

Statistical Data ◽

Activation Function ◽

Decision Making Process ◽

Neural Network Architecture ◽

Acceptable Accuracy ◽

The Neural Network ◽

Sigmoid Activation Function

The problem of applying neural networks to calculate ratings used in banking in the decision-making process on granting or not granting loans to borrowers is considered. The task is to determine the rating function of the borrower based on a set of statistical data on the effectiveness of loans provided by the bank. When constructing a regression model to calculate the rating function, it is necessary to know its general form. If so, the task is to calculate the parameters that are included in the expression for the rating function. In contrast to this approach, in the case of using neural networks, there is no need to specify the general form for the rating function. Instead, certain neural network architecture is chosen and parameters are calculated for it on the basis of statistical data. Importantly, the same neural network architecture can be used to process different sets of statistical data. The disadvantages of using neural networks include the need to calculate a large number of parameters. There is also no universal algorithm that would determine the optimal neural network architecture. As an example of the use of neural networks to determine the borrower's rating, a model system is considered, in which the borrower's rating is determined by a known non-analytical rating function. A neural network with two inner layers, which contain, respectively, three and two neurons and have a sigmoid activation function, is used for modeling. It is shown that the use of the neural network allows restoring the borrower's rating function with quite acceptable accuracy.

Download Full-text