Neural Network Inverse Model Research Articles

PEM Fuel Cell Voltage Neural Control Based on Hydrogen Pressure Regulation

Fuel cells are promising devices to transform chemical energy into electricity; their behavior is described by principles of electrochemistry and thermodynamics, which are often difficult to model mathematically. One alternative to overcome this issue is the use of modeling methods based on artificial intelligence techniques. In this paper is proposed a hybrid scheme to model and control fuel cell systems using neural networks. Several feature selection algorithms were tested for dimensionality reduction, aiming to eliminate non-significant variables with respect to the control objective. Principal component analysis (PCA) obtained better results than other algorithms. Based on these variables, an inverse neural network model was developed to emulate and control the fuel cell output voltage under transient conditions. The results showed that fuel cell performance does not only depend on the supply of the reactants. A single neuro-proportional–integral–derivative (neuro-PID) controller is not able to stabilize the output voltage without the support of an inverse model control that includes the impact of the other variables on the fuel cell performance. This practical data-driven approach is reliably able to reduce the cost of the control system by the elimination of non-significant measures.

A hybrid approach based on inverse neural network to determine optimal level of energy consumption in electrical power generation

Power plants are a strategic infrastructure industry in each country and provide a strong driving force for the development of other industries. Hence, the performance of a power plant should be optimized by adjusting input parameters to maximize efficiency. This paper mainly proposes a novel hybrid approach that aims to present an optimal level of input parameters of a power plant integrating Data Envelopment Analysis (DEA), Artificial Neural Network and Inverse Problem in a new way. The inputs for a power plant system include internal electricity, fossil fuel and water consumption which are used to generate electrical power. First, the collected data are evaluated using a DEA model to identify and determine the days that operated efficiently. Then, the efficient vectors including resources consumption and generated power in the days evaluated in the previous stage are entered into the artificial neural network. Afterward, the weights and biases of the trained network are extracted for use in the inverse neural network. The inverse neural network model is presented as a nonlinear optimization problem with Gaussian constraints. Because of the severely ill-posed nature of the problem, to solve and obtain reasonable solutions, an innovative approach is used to determine the optimal level of inputs in less time and more accuracy. Based on this approach, for solving the nonlinear model and reducing the number of candidate solutions, the optimal interval is computed for each input and each day. A real case study of the thermal power plant in Iran is presented to show the abilities of the proposed approach.

RETRACTED ARTICLE: Research on hot-rolling steel products quality control based on BP neural network inverse model

Taking the hot-rolling products made by an iron and steel company as the research object, this paper builds the inverse model reflecting the relationship between the hot-rolling steel product performance indicators, the chemical composition of steel and the rolling technological parameters by using the BP neural network. So the purpose of getting technological parameters is achieved, according to the given steel performance indicators. Combining the BP neural network, adaptive inverse control with internal model control theory, this paper builds the BP neural network inverse model with multiple input and single output based on internal model control. Therefore, it realizes the inverse mapping between the output and the input variables of the BP neural network. And the output variables can be obtained according to the input variables. Besides, this paper also gives the detailed steps to solve the inverse model. Then, the model is applied to the hot-rolling steel products quality control system. The performance indicators of the hot-rolling products are set up, and the rolling technological parameters—the rolling crimp temperature—are solved. The model realizes the controllability of rolling technological parameters. Finally, through the verification of hot-rolling products quality control positive system, the error is in line with the enterprise production requirements.

Selection of the Informative Input Parameters for the Inverse Neural-Network Models of Observed Systems

We consider the problem of determination of the parameters of a measurement grid, which guarantees the exactness and stability of the solution of the inverse problem. The choice of the points of measurements is performed under the assumption of existence of the most informative data. We present the results illustrating the influence of the number of measurement points on the data of reconstruction of the parameters of the load function acting upon the cylindrical shell in a strip located along the length of the shell.

Decoupling control of a five-phase fault-tolerant permanent magnet motor by radial basis function neural network inverse

This paper proposes a new decoupled control for a five-phase in-wheel fault-tolerant permanent magnet (IW-FTPM) motor drive, in which radial basis function neural network inverse (RBF-NNI) and internal model control (IMC) are combined. The RBF-NNI system is introduced into original system to construct a pseudo-linear system, and IMC is used as a robust controller. Hence, the newly proposed control system incorporates the merits of the IMC and RBF-NNI methods. In order to verify the proposed strategy, an IW-FTPM motor drive is designed based on dSPACE real-time control platform. Then, the experimental results are offered to verify that the d-axis current and the rotor speed are successfully decoupled. Besides, the proposed motor drive exhibits strong robustness even under load torque disturbance.

A Low-Cost Maximum Power Point Tracking System Based on Neural Network Inverse Model Controller

This work presents the design, modeling, and implementation of a neural network inverse model controller for tracking the maximum power point of a photovoltaic (PV) module. A nonlinear autoregressive network with exogenous inputs (NARX) was implemented in a serial-parallel architecture. The PV module mathematical modeling was developed, a buck converter was designed to operate in the continuous conduction mode with a switching frequency of 20 KHz, and the dynamic neural controller was designed using the Neural Network Toolbox from Matlab/Simulink (MathWorks, Natick, MA, USA), and it was implemented on an open-hardware Arduino Mega board. To obtain the reference signals for the NARX and determine the 65 W PV module behavior, a system made of a 0.8 W PV cell, a temperature sensor, a voltage sensor and a static neural network, was used. To evaluate performance a comparison with the P&O traditional algorithm was done in terms of response time and oscillations around the operating point. Simulation results demonstrated the superiority of neural controller over the P&O. Implementation results showed that approximately the same power is obtained with both controllers, but the P&O controller presents oscillations between 7 W and 10 W, in contrast to the inverse controller, which had oscillations between 1 W and 2 W.

Fuzzy sliding mode controller design for semi-active seat suspension with neuro-inverse dynamics approximation for MR damper

To improve the ride comfort of car, this paper proposed a semi-active seat suspension with magneto-rheological (MR) damper and designed a new fuzzy sliding mode controller with expansion factor (FSMCEF) based on the neuro-inverse dynamics approximation of the MR damper. This FSMCEF combines the advantages of both sliding mode controller (SMC) and fuzzy controller (FC) with expansion factor (EF), and it takes an ideal skyhook model as the reference, and creates a sliding mode control law based on the errors dynamics between the seat suspension and its reference model. Further fuzzy rules are used to suppress the chattering occurred in the above sliding mode control by fuzzifying the sliding mode surface and its derivative. Moreover, in order to compute the required control current for MR damper after solving the desired control force using FSMCEF, this paper presented a BP algorithm based neural network inverse model, located between the FSMCEF and the MR damper, taking the displacement, velocity of the MR damper and the desired control force output by FSMCEF as its input, and predicting the control current required to input MR damper. The predicting error and stability of the neural network inverse model for MR is investigated by sample testing. In addition, the stability analysis of FSMCEF is also completed by under nominal system and non-nominal system with parameter uncertainty and external disturbance. The results of numerical simulations show that the vibration reduction effect of the semi-active seat is obviously improved using FSMCEF compared with using PID controller and SMC.

Modeling and optimization of a solar parabolic trough concentrator system using inverse artificial neural network

In this paper, an artificial neural network inverse (ANNi) model is applied to optimize the thermal performance (η) of parabolic trough concentrators. A feedforward neural network architecture is trained using an experimental database from parabolic trough concentrators operations. Rim angle (φr), inlet (Tin) and outlet (Tout) fluid temperatures, ambient temperature (Ta), water flow (Fw), direct solar radiation (Gb), and the wind velocity (Vw) were used as main input variables within the neural network model to estimate the thermal performance with a correlation coefficient of R2 = 0.9996 between experimental and simulated values. The sensitivity analysis is carried out to verify the effect of all input variables. The optimal operation conditions of parabolic trough concentrators are established using artificial neural network inverse modeling (ANNi) to achieve optimal operation conditions of parabolic trough concentrators. The results indicated that ANNi is a feasible tool for Parabolic Trough Concentrators optimization.

Offset-free adaptive nonlinear model predictive control with disturbance observer for DC-DC buck converters

The aim of this paper is to design a nonlinear model predictive control for DC-DC buck converters to track constant reference signals with zero steady-state error. The online trained neural network (NN) model is employed as the predictor and the steady-state error, which is called the offset, is studied in the presence of the changes in system parameters and the external disturbances. The stability of the closed-loop system is investigated using the Lyapunov direct theory. The proposed method can provide offset-free behavior in the presence of constant disturbances. For rejecting nonconstant disturbances, a nonlinear disturbance observer based on the NN inverse model is proposed. Due to wide applications of the DC-DC converter in power electronics, control of its output voltage is considered in this paper. The effectiveness of the proposed control method is demonstrated by experimental results.

Neural network approach to reduce dynamic measurement errors

<p>The neural network inverse model of a sensor with filtration of the sequentially recovered signal is considered. This model effectively reduces the dynamic measurement errors due to deep mathematical processing of measurement data. The result of the experimental data processing of a dynamic temperature measurement validates the efficiency of the proposed neural network approach to reduce dynamic measurement errors.</p>

INTELLIGENT CONTROLLERS FOR VELOCITY TRACKING OF TWO WHEELED INVERTED PENDULUM MOBILE ROBOT

Velocity tracking is one of the important objectives of vehicle, machines and mobile robots. A two wheeled inverted pendulum (TWIP) is a class of mobile robot that is open loop unstable with high nonlinearities which makes it difficult to control its velocity because of its nature of pitch falling if left unattended. In this work, three soft computing techniques were proposed to track a desired velocity of the TWIP. Fuzzy Logic Control (FLC), Neural Network Inverse Model control (NN) and an Adaptive Neuro-Fuzzy Inference System (ANFIS) were designed and simulated on the TWIP model. All the three controllers have shown practically good performance in tracking the desired speed and keeping the robot in upright position and ANFIS has shown slightly better performance than FLC, while NN consumes more energy.

Intelligent coordinated controller design for a 600 MW supercritical boiler unit based on expanded-structure neural network inverse models

Under present widespread automatic generation control (AGC) centered on regional power grid, a large-capacity coal-fired supercritical (SC) power unit often operates under wide-range variable load conditions. Since a SC once-through boiler unit is represented by a typical multivariable system with large inertia and non-linear, slow time-variant and time-delay characteristics, it often makes the coordinated control quality deteriorate under wide-range loading conditions, and thus influences the unit load response speed and leads to heavy fluctuation of the main steam pressure. To improve the SC unit’s coordinated control quality with advanced intelligent control strategy, the neural-network (NN) based expanded-structure inverse system models of a 600MW SC boiler unit were investigated. A feedforward neural network with time-delayed inputs and time-delayed output feedbacks was adopted to establish the inverse models for the load and the main steam pressure characteristics. Based on the model, a neural network inverse coordinated control scheme was designed and tested in a full-scope power plant simulator of the given SC power unit, which showed that the proposed coordinated control scheme can achieve better control results compared to the original PID coordinated control.

A new model predictive control scheme for energy and cost savings in commercial buildings: An airport terminal building case study

Predictive control technology for heating, ventilation and air conditioning (HVAC) systems has been proven to be an effective way to reduce energy consumption and improve thermal comfort within buildings. Such methods rely on models to accurately predict the thermal dynamics of a specific building to achieve the optimal control. Implementing a predictive control at the building level faces several challenges, since buildings’ thermal dynamics are nonlinear, time-varying, and contain several uncertainties. This paper presents a hybrid model predictive control (HMPC) scheme, which can minimise the energy and cost of running HVAC systems in commercial buildings. The proposed control framework combines a classical MPC with a neural network feedback linearisation method. The control model for the HMPC is developed using a simplified physical model, while the nonlinearity associated with HVAC process is handled independently by an inverse neural network model. To achieve the maximum energy saving, the proposed MPC integrates several advanced air-conditioning control strategies, such as an economizer control, an optimal start-stop control, and a load shifting control. This approach has been tested at the check-in hall of the T-1 building of the Adelaide Airport, through simulations and a field experiment. The merits of the proposed method compared to the existing control method are analysed from both the energy saving and cost saving points of view. The result shows that the proposed HMPC scheme performs reasonably well, and achieves a considerable amount of savings without violating thermal comfort.

Approaches in a Sensor Model of Error Correction in Dynamic Measurements

The article describes three approaches in error correction of dynamic measurements. A sensor model is used in these approaches. The first one is based on the modal control of the dynamic behavior and adapting parameters of a sensor model by direct search. We propose a method of error evaluation in dynamic measurements with a priori information on characteristics of measured signal and noise of the sensor available. The second approach concerns the neural network representation of a sensor. Neural network inverse sensor model and the algorithm for its training by minimizing mean-squared dynamic measurements error criterion are proposed. The third approach is based on the introduction of sliding mode control into the measuring system with modal control of dynamic behavior to achieve the similarity of the sensor model output to the sensor output.

Simple adaptive control of seismically excited structures with MR dampers

In this paper, Simple Adaptive Control (SAC) method is used to mitigate the detrimental effects of earthquakes on MR-damper equipped structures. Acceleration Feedback (AF) is utilized since measuring the acceleration response of structures is known to be reliable and inexpensive. The SAC is simple, fast and as an adaptive control scheme, is immune against the effects of plant and environmental uncertainties. In the present study, in order to translate the desired control force into an applicable MR damper command voltage, a neural network inverse model is trained, validated and used through the simulations. The effectiveness of the proposed AF-based SAC control system is compared with optimal H2/LQG controllers through numerical investigation of a three-story model building. The results indicate that the SAC controller is substantially effective and reliable in both undamaged and damaged structural states, specifically in reducing acceleration responses of seismically excited buildings.

Delay Separated Neural Network Inverse Control in Main-Steam Temperature System

In order to improve the control effect of the main steam temperature with large time delay, this paper proposed a delay separated neural network inverse (DSNNI) control scheme. The delay time and the positive model without delay were given by using adaptive linear element and BP network. The neural network inverse model of the positive model without delay was built on that basis. An appropriate reference model was selected to make the inverse model’s output smoothing. It is an open-loop control system when the model is cascaded into original system. It will avoid the instability caused by the closed-loop control systems. Off-line identification and on-line identification are combined to get the inverse model in order to reduce the steady-state error and make the system have fine adaptive capacity. Detail simulation tests are carried out on the given 300MW power unit. Tests show that the neural network inverse control with delay time separation can get rapid and smooth output for the main steam temperature system. It is able to overcome the adverse effects caused by the time delay and the parameters changes. Compared with the cascade PID controller, the adjustment time of DSNNI reduces from 600s to380s and shows faster response, better robustness and anti-interference performance. Full Text: PDF DOI: http://dx.doi.org/10.11591/telkomnika.v12i7.4696

Neural Network Inverse Model Control Strategy: Discrete-Time Stability Analysis for Relative Order Two Systems

This paper discusses the discrete-time stability analysis of a neural network inverse model control strategy for a relative order two nonlinear system. The analysis is done by representing the closed loop system in state space format and then analyzing the time derivative of the state trajectory using Lyapunov’s direct method. The analysis shows that the tracking output error of the states is confined to a ball in the neighborhood of the equilibrium point where the size of the ball is partly dependent on the accuracy of the neural network model acting as the controller. Simulation studies on the two-tank-in-series system were done to complement the stability analysis and to demonstrate some salient results of the study.

Modeling and inverse feedforward control for conducting polymer actuators with hysteresis

Conducting polymer actuators are biocompatible with a small footprint, and operate in air or liquid media under low actuation voltages. This makes them excellent actuators for macro- and micro-manipulation devices, however, their positioning ability or accuracy is adversely affected by their hysteresis non-linearity under open-loop control strategies. In this paper, we establish a hysteresis model for conducting polymer actuators, based on a rate-independent hysteresis model known as the Duhem model. The hysteresis model is experimentally identified and integrated with the linear dynamics of the actuator. This combined model is inverted to control the displacement of the tri-layer actuators considered in this study, without using any external feedback. The inversion requires an inverse hysteresis model which was experimentally identified using an inverse neural network model. Experimental results show that the position tracking errors are reduced by more than 50% when the hysteresis inverse model is incorporated into an inversion-based feedforward controller, indicating the potential of the proposed method in enabling wider use of such smart actuators.

The control of MSF desalination plants based on inverse model control by neural network

In this paper, a nonlinear inverse model control strategy based on neural network is proposed for MSF desalination plant. Artificial neural networks (ANNs) can handle complex and nonlinear process relationships, and are robust to noisy data. The designed neural networks consist of three layers identified from input–output data and trained with a descent gradient algorithm. The set point tracking performance of the proposed method was studied when the disturbance is present in the MSF system. Three controllers are designed for controlling the top brine temperature, the level of last stage and salinity. These results show that a neural network inverse model control strategy (NNINVMC) is robust and highly promising to be implemented in such nonlinear systems. Also the comparison between the top brine temperature of the proposed model and NN predicted data from the literature supports the accuracy of the model.

Engineering Neutron Diffraction Data Analysis with <i>Inverse</i> Neural Network Modeling

Integration of engineering neutron diffraction data analysis and solid mechanics modeling is a powerful method to deduce in-situ constitutive behavior of materials. Since diffraction data originates from spatially discrete subsets of the material volume, extrapolation of the data to the behavior of the overall sample is non-trivial. The finite element model has been widely used for interpreting diffraction data by optimizing model parameters via iterative processes. In order to maximize the rigor of such analysis and to increase fitting efficiency and accuracy, we have developed an optimization algorithm based on the neural network architecture. The inverse neural network model reveals parameter sensitivity quantitatively during a training process, and yields more accurate phase specific constitutive laws of the composite materials compared to the conventional method once networks are successfully trained.