Ferenc Szeifert - Academia.edu (original) (raw)

Papers by Ferenc Szeifert

OTKA 49534 Zárójelentés, 2008

A kutatás során egy termék- és technológiafejlesztési keretrendszer került kidolgozásra a tágabb ... more A kutatás során egy termék- és technológiafejlesztési keretrendszer került kidolgozásra a tágabb értelemben vett kísérlettervezési és adatelemzési technikák célirányos fejlesztése alapján. A kidolgozott technikák közös jellemzője a technológia fejlesztés lépései során keletkező információk hatékony kezelése. A technológia üzemeltetése során keletkező adatokból a potenciálisan hasznos információk feltárására adatbányászati technikákat alkalmaztunk, melyeket a vegyészmérnöki, és természettudományos ismeretek hatékony kezelése érdekében a legkorszerűbb folyamat szimulációs eszközökkel integráltunk. Az eredményeket egy a Springer leányvállalatánál (Birkhauser) megjelent monográfiában, egy magyar nyelvű (tan)könyvben, és tíznél több nemzetközi referált folyóiratban publikáltuk. A kidolgozott keretrendszer ipari alkalmazása folyamatban van. | A new framework for product and process development has been worked out based on the research of novel data mining and experiment design algorithms. The common feature of the proposed tools and methodologies is the effective management of the information through the whole life-cycle of process development. In the proposed framework data mining techniques have been applied to extract potentially useful information from historical process data. The related process-relevant a priori information is incorporated by advanced simulation techniqes. The resullts has been published in a reserch monogaph by Biskhauser and in more than ten peer-reviewed journals. The industrial application of the proposed framework has been also studied.

Selecting the order of an input-output model of a dynamical system is a key step toward the goal ... more Selecting the order of an input-output model of a dynamical system is a key step toward the goal of system identication. By determining the smallest regression vector di- mension that allows accurate prediction of the output, the false nearest neighbors algorithm (FNN) is a useful tool for linear and also for nonlinear systems. The one parameter that needs to be

Pattern Recognition Letters, 2003

Journal of Process Control, 2004

Industrial & Engineering Chemistry Research, 1999

Engineering Applications of Artificial Intelligence, 2005

Computers & Chemical Engineering, 2005

Mathematics and Computers in Simulation, vol. 51, num. 1, pp. 119-132, 1999

This paper proposes a direct adaptive fuzzy-model-based control algorithm. The controller is base... more This paper proposes a direct adaptive fuzzy-model-based control algorithm. The controller is based on an inverse semi-linguistic fuzzy process model, identified and adapted via input-matching technique. For the adaptation of the fuzzy model a general learning rule has been developed employing gradient-descent algorithm. The on-line learning ability of the fuzzy model allows the controller to be used in applications, where the knowledge to control the process does not exist or the process is subject to changes in its dynamic characteristics. To demonstrate the applicability of the method, a realistic simulation experiments were performed for a non-linear liquid level process. The proposed direct adaptive fuzzy logic controller is shown to be capable of handling non-linear and time-varying systems dynamics, providing good overall system performance.

Computational Intelligence for Modelling, Control and Automation, pp. 265-270., 1999

This paper presents a new fuzzy model based predictive controlalgorithm. The proposed predictor c... more This paper presents a new fuzzy model based predictive controlalgorithm. The proposed predictor corrector controller is based on a hybrid fuzzymodel, which consists of a fuzzy steady state model and a gain independent impulseresponse model. The real-time control of a laboratory-sized heating system is chosenas a nonlinear case study for the demonstration of the proposed control algorithm.The results show that the proposed algorithm is capable of controlling the nonlinear process that operates over wide range.

In order to solve the problem of model based control arising from the process model has to be obt... more In order to solve the problem of model based control arising from the process model has to be obtained by using small amount and different type of available information, a fuzzy modeling framework has been developed for the utilization of a priori knowledge. The proposed modeling approach transforms the different types of information into the structure of the model (fuzzy rule base), constraints defined on the parameters and variables, dynamic local model or data, and steady-state data or model. This modeling step is followed by an optimization procedure based on these transformed information. The paper describes one element of this framework that was developed to use prior knowledge in constrained adaptation of the rule consequences of Takagi-Sugeno fuzzy models. Experimental results have been obtained for a laboratory setup consisting of two cascaded tanks. It has been shown that by using constrained adaptation, good control performance can be achieved for a nonlinear, time-varying process. 1

INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH., vol. 42, num. 17, pp. 4043-4049, 2003

Data smoothening and re-sampling are often necessary to handle data obtained from laboratory and ... more Data smoothening and re-sampling are often necessary to handle data obtained from laboratory and industrial experiments. This paper presents a new algorithm for incorporating prior knowledge into spline-smoothing of interrelated multivariate data. Prior knowledge based on the visual inspection of the variables and/or knowledge about the assumed balance equations can be transformed into linear equality and inequality constraints on the parameters of the splines. The splines than can be simultaneously identified from the available data by solving one quadratic programming problem. To demonstrate the applicability of the method two examples are given. In the first example, the proposed approach has been applied to the identification of kinetic parameters of a simulated reaction network, while in the second example data taken from an industrial batch reactor is analyzed. The results show that, when the proposed constrained spline-smoothing algorithm is applied, not only better fitting to the data points is achieved, but also the performance of the estimation of the kinetic parameters improves with regard to the case where no prior knowledge is involved.

Computer Aided Chemical Engineering; vol. 24, pp. 751-756, Jan 2007

This work focuses on runaway behaviour of industrial catalytic tube-reactor and presents how deci... more This work focuses on runaway behaviour of industrial catalytic tube-reactor and presents how decision trees can be used for forecasting the runaway. The steady-state simulator of the reactor and runaway criterion based on Ljapunov's indirect stability analysis have been used to generate the database used by the decision tree induction algorithm. The extracted logical rules can be used in an operator support system (OSS), and are useful for working out safe operating strategies.

INDUSTRIAL & ENGINEERING CHEMISTRY RESEARCH; vol. 49, num. 2, pp. 658-668, 2009

The operation of complex production processes is one of the most important research and developme... more The operation of complex production processes is one of the most important research and development problems in process engineering. A safety instrumented system (SIS) performs specified functions to achieve or maintain a safe state of the process when unacceptable or dangerous process conditions are detected. A logic solver is required to receive the sensor input signal(s), to make appropriate decisions based on the nature of the signal(s), and to change its outputs according to user-defined logic. The change of the logic solver output(s) results in the final element(s) taking action on the process (e.g., closing a valve) to bring it back to a safe state. Alarm management is a powerful tool to support the operators' work to control the process in safe operating regions and to detect process malfunctions. Predictive alarm management (PAM) systems should be able not only to detect a dangerous situation early enough, but also to give advice to process operators which safety action (or safety element(s)) must be applied. The aim of this paper is to develop a novel methodology to support the operators how to make necessary adjustments in operating variables at the proper time. The essential of the proposed methodology is the simulation of the effect of safety elements over a prediction horizon. Since different manipulations have different time demand to avoid the evolution of the unsafe situation (safety time), the process operators should know which safety action(s) should be taken at a given time. For this purpose a method for model based predictive stability analysis has been worked out based on Lyapunov's stability analysis of simulated state trajectories. The proposed algorithm can be applied to explore the stable and unstable operating regimes of a process (set of safe states), information that can be used for PAM. The developed methodology has been applied to two industrial benchmark problems related to the thermal runaway.

Acta Polytechnica Hungarica; vol. 4, num. 4, pp. 121-131, 2007

Safe and optimal operation of complex production processes is one of the mostimportant research a... more Safe and optimal operation of complex production processes is one of the mostimportant research and development problems in process engineering. This problem is themost relevant at the design of the optimal feeding profile of fed-batch chemical reactorsdue to the nonlinear and unstable dynamical behavior of the processes. This paper showsthat how the optimal feeding policy can be determined in fed-batch reactors by sequentialquadratic programming, classical evolutionary strategy (ES) and the advanced version ofES that is based on covariance matrix adaptation. A multi-objective function was createdand the search space was constrained in case of all of the three applied algorithms. Theswitching times between states in the feeding trajectory and the feed rates in each statewere manipulated to find the global minima of the objective function. To obtain the optimalfeeding policy the first-principle model of a pilot fed-batch reactor was implemented inMATLAB and applied as a dynamic simulator of the process. Off-line optimization processwas carried out in case of different dosing time distribution. As the results show asignificant improvement can be achieved in process performance applying advanced ESbased optimization algorithms to generate feeding trajectories.

Industrial & Engineering Chemistry Research, vol. 44, num. 9, pp. 3178-3186, 2005

Linear-in-parameters models are quite widespread in process engineering, e.g., nonlinear additive... more Linear-in-parameters models are quite widespread in process engineering, e.g., nonlinear additive autoregressive models, polynomial ARMA models, etc. This paper proposes a new method for the structure selection of these models. The method uses genetic programming to generate nonlinear input−output models of dynamical systems that are represented in a tree structure. The main idea of the paper is to apply the orthogonal least squares (OLS) algorithm to estimate the contribution of the branches of the tree to the accuracy of the model. This method results in more robust and interpretable models. The proposed approach has been implemented as a freely available MATLAB Toolbox, www.fmt.veim.hu/softcomp. The simulation results show that the developed tool provides an efficient and fast method for determining the order and structure for nonlinear input−output models.

5th International Conference on Intelligent Systems Design and Applications, Volume 2005, 2005, , Pages 314-319, 2005

It is often desirable to simultaneously handle several objectives and constraints in practical op... more It is often desirable to simultaneously handle several objectives and constraints in practical optimization problems. In some cases, these objectives and constraints are non-commensurable and they are not explicitly/mathematically available. For this kind of problems, interactive optimization may be a good approach. Interactive optimization means that a human user evaluates the potential solutions in qualitative way. In recent years evolutionary computation (EC) was applied for interactive optimization, which approach has became known as interactive evolutionary computation (IEC). The aim of this paper is to propose a new interactive optimization method based on particle swarm optimization (PSO). PSO is a relatively new population based optimization approach, whose concept originates from the simulation of simplified social systems. The paper shows that interactive PSO cannot be based on the same concept as IEC because the information sharing mechanism of PSO significantly differs from EC. So this paper proposes an approach which considers the unique attributes of PSO. The proposed algorithm has been implemented in MATLAB (IPSO toolbox) and applied to a case-study of temperature profile design of a batch beer fermenter. The results show that IPSO is an efficient and comfortable interactive optimization algorithm. The developed IPSO toolbox (for Mat-lab) can be downloaded from the Web site of the authors: http://www.fmt.vein.hu/softcomp/ipso.

International Journal of Systems Science, vol. 31, num. 4, pp. 457-466, 2000

This paper looks at a new method of modelling non-linear dynamic processes, using grid-type Sugen... more This paper looks at a new method of modelling non-linear dynamic processes, using grid-type Sugeno fuzzy models and a priori knowledge. The proposed hybrid fuzzy convolution dynamic model consists of a non-linear fuzzy steady-state static, and a gain-independent impulse response model-based dynamic part. The modelling of non-linear pH processes is chosen as a realistic case study for the demonstration of the proposed modelling approach. The off-line identified hybrid fuzzy convolution model is shown to be capable of modelling the non-linear process and providing better multi-step prediction than the conventional grid-type Sugeno fuzzy model.

IEEE International Conference on Fuzzy Systems, Volume 2, Pages 951-956, 1999

This paper presents an algorithm for incorporating of a priori knowledge into data-driven identif... more This paper presents an algorithm for incorporating of a priori knowledge into data-driven identification for dynamic fuzzy models of the Takagi-Sugeno type. Knowledge about the modeled process such as its stability minimal or maximal static gain, or the settling time of its step response can be translated into inequality constraints on the consequent parameters. By using input-output data, optimal parameter values are then found by means of quadratic programming. The proposed approach was successfully applied to the identification of a laboratory liquid level process

Chemical Engineering Research & Design; vol. 77, num. 4, pp. 281-290, 1999

This study presents an adaptation method for Sugeno fuzzy inference systems that maintain the rea... more This study presents an adaptation method for Sugeno fuzzy inference systems that maintain the readability and interpretability of the fuzzy model during and after the learning process. This approach can be used for the modelling of dynamical systems and for building adaptive model-based control algorithms for chemical processes.

The gradient-descent based learning algorithm can be used on-line to form an adaptive fuzzy controller—this ability allows these controllers to be used in applications where the knowledge to control the process does not exist or the process is subject to changes in its dynamic characteristics. The proposed approach was applied in an internal model (IMC) fuzzy control structure based on the inversion of the fuzzy model. The adaptive fuzzy controller was applied in the control of a non-linear plant and is shown to be capable of providing good overall system performance.