scispace - formally typeset
Proceedings ArticleDOI: 10.1109/IC4.2015.7375581

Artificial neural network based inverse model control of a nonlinear process

01 Sep 2015-pp 1-6
Abstract: In process industries the non linear process control is a challenging and difficult task due to its non linear behavior, delays and time variation between inputs and outputs of system. Conical tank system is one such non linear process which is widely used in process industries due of its non linear shape and easy flow of liquid across its cross section area. As conical tank is inherently non linear it becomes difficult to model the linear plant equation for the same. The control of liquid level in conical tank is a complex and complicated task because of its constantly changing cross section area. So, Artificial Neural network (ANN) based controller is designed because of its ability to model non linear systems and its inverses. The Direct Inverse Control (DIC) designed using ANN is mainly dependent on the inverse response of the system which is difficult task to obtain it analytically. In this paper, ANN based DIC is trained by Levenberg Marquardt Back propagation algorithm and helps to obtain optimized response/performance of the system. The simulation results show that direct inverse control realize a good dynamic behaviour of interacting and non interacting conical tank system.

...read more

Topics: Nonlinear system (52%), Control theory (52%), Process control (52%) ...read more
Citations
  More

Journal ArticleDOI: 10.1007/S13369-017-3034-9
Abstract: This paper performs the comparative study of two feed-forward neural networks: radial basis function network (RBFN), multilayer feed-forward neural network (MLFFNN) and a recurrent neural network: nonlinear auto-regressive with exogenous inputs (NARX) neural network for their ability to provide an adaptive control of nonlinear systems. Dynamic back-propagation algorithm is used to derive parameter update equations. To ensure stability and faster convergence, an adaptive learning rate is developed in the sense of discrete Lyapunov stability method. Both parameter variation and disturbance signal cases are considered for checking and comparing the robustness of controller. Three simulation examples are considered for carrying out this study. The results so obtained reveal that RBFN-based controller is performing better than that of NARX- and MLFFNN-based controllers.

...read more

Topics: Recurrent neural network (63%), Adaptive control (61%), Artificial neural network (59%) ...read more

5 Citations


Proceedings ArticleDOI: 10.1109/ICIRD47319.2019.9074761
28 Jun 2019-
Abstract: This paper discusses about designing a direct inverse control based back-propagation neural network for a skid steer model boat. The boat is modeled into a MIMO system, with port side and starboard propeller as input, while yaw, surge velocity and sway velocity as output. The inverse plant model was created for the controller and tested with an identification model of the plant. The simulation result is the plant output follows the desired output value that is fed to the inverse controller with the normalized mean sum square error for yaw 0.1203, surge velocity 0.4459 and sway velocity 0.1723.

...read more

Topics: Skid (automobile) (52%)

1 Citations


Journal ArticleDOI: 10.1177/0959651820988189
27 Jan 2021-
Abstract: In this study, an adaptive control based on fuzzy adapting rate for neural emulator of nonlinear systems having unknown dynamics is proposed. The indirect adaptive control scheme is composed by the...

...read more

1 Citations


Open accessPosted Content
08 Apr 2021-arXiv: Learning
Abstract: This paper presents a novel Generative Neural Network Architecture for modelling the inverse function of an Artificial Neural Network (ANN) either completely or partially. Modelling the complete inverse function of an ANN involves generating the values of all features that corresponds to a desired output. On the other hand, partially modelling the inverse function means generating the values of a subset of features and fixing the remaining feature values. The feature set generation is a critical step for artificial neural networks, useful in several practical applications in engineering and science. The proposed Oracle Guided Generative Neural Network, dubbed as OGGN, is flexible to handle a variety of feature generation problems. In general, an ANN is able to predict the target values based on given feature vectors. The OGGN architecture enables to generate feature vectors given the predetermined target values of an ANN. When generated feature vectors are fed to the forward ANN, the target value predicted by ANN will be close to the predetermined target values. Therefore, the OGGN architecture is able to map, inverse function of the function represented by forward ANN. Besides, there is another important contribution of this work. This paper also introduces a new class of functions, defined as constraint functions. The constraint functions enable a neural network to investigate a given local space for a longer period of time. Thus, enabling to find a local optimum of the loss function apart from just being able to find the global optimum. OGGN can also be adapted to solve a system of polynomial equations in many variables. The experiments on synthetic datasets validate the effectiveness of OGGN on various use cases.

...read more

Topics: Feature vector (57%), Artificial neural network (55%), Inverse function (55%) ...read more

Book ChapterDOI: 10.1007/978-3-030-63322-6_84
14 Oct 2020-
Abstract: In this contribution, the temperature control of the 3D printer heatbed is observed. As the heat exchange power is strictly limited and the thermal process time constants are naturally around tens and hundreds of seconds, these processes are basically slow. The measuring of new data is time consuming, which can cause the profit loss in case of experiments in the production. Moreover, the finding of better control method can lead to significant monetary savings. One of the scopes of this article is to find out if it’s possible to built-up the neural network-based controller system together with the internal model control strategy providing better performance with data obtained in the production, where simply tuned PSD controller was used. The suitable order of the heating system is observed together with the size of the sampling period and neural network topology. The controllability with best performing neural networks is verified on the 3D printer heating bed.

...read more

Topics: Control theory (56%), Temperature control (54%), Controllability (52%) ...read more
References
  More

Open accessJournal Article
Abstract: This paper describes a nonlinear model of conical tank level control system and real time system designs are analysed and their implementation in SIMULINK is outlined. Level control of a conical tank is a complex issue because of the nonlinear nature of the tank. For each stable operating point, a First Order Process model was identified using process reaction curve method; the Control is done and comparison of the synthesis method and skogestad method is clarified.

...read more

Topics: Operating point (51%), Nonlinear system (50%)

32 Citations


Open accessJournal ArticleDOI: 10.5120/14891-3358
Abstract: The control of liquid level is mandatory in process industries. But the control of nonlinear process is complex. Many process industries use conical tanks because of its non linear shape which contributes better drainage for solid mixtures, slurries and viscous liquids. So, level control of conical tank presents a challenging task due to its nonlinearity and constantly changing cross-section. The main objective is to implement the suitable controller design for conical tank system to maintain the desired level. In this paper it is proposed to obtain the mathematical modelling of a conical tank system and to design model based controller (Internal Model Control) for controlling the level in it. The controller will be simulated using MATLAB SIMULINK. By using the advanced control scheme it is expected to have better closed loop performance and robustness when compared to PID controller.

...read more

Topics: Conical surface (55%)

19 Citations


Open accessJournal Article
Abstract: Non-linear process control is a difficult problems in process industries. Conical tank level control is one among them. Real processes often exhibits nonlinear behavior, time variance and delays between inputs and outputs. Conical tanks are widely used in many industries due to its shape which provides easy discharge of water when compared to other tanks.. Moreover, liquid level control of a conical tank is still challenging for typical process control because of its nonlinearities by a reason of constantly changing cross section area. In this paper the mathematical modeling of three tank conical interacting and non-interacting system is designed by Wiener model PI controller(WMPI) where the tuning rules based on Chidambaram method and the performance criteria are related with Internal mode controller(IMC). Also in this paper we analyze dynamic behavior among interacting and non-interacting system.

...read more

Topics: Control theory (55%), Process control (53%), PID controller (52%)

7 Citations


Open access
01 Jan 2014-
Topics: Conical surface (54%)

4 Citations


Proceedings ArticleDOI: 10.1109/ICACCCT.2012.6320799
V.R. Ravi, T. Thyagarajan1, B. Puviyarasi2Institutions (2)
04 Oct 2012-
Abstract: The implementation of control algorithms for MIMO systems is often complicated due to variations in process dynamics that occur because of changes in operating point and the characteristics of non linear dynamic coupling. Such changes often render the performance of existing decoupled based decentralized fixed gain PID and Gain scheduling PID controllers unsatisfactory. The dynamic uncertainty associated with MIMO systems make existing model based decoupling impractical for real time system. This work presents control of non linear Two Conical Tank Interacting Level System (TCTILS) using centralised Neuro controller. The TCTILS is considered as two tank benchmark problem used by many researchers. Simulation results show that the centralised Neuro controller realise a good dynamic behaviour of the TCTILS, a perfect level tracking with lesser overshoot, lesser settling time, reduced interaction and good rejection of external load disturbances.

...read more

Topics: Gain scheduling (58%), PID controller (58%), Operating point (51%)

4 Citations


Performance
Metrics
No. of citations received by the Paper in previous years
YearCitations
20212
20201
20191
20181