S
Journal of Computer Science and Control Systems http://electroinf.uoradea.ro/index.php/reviste/jcscs.html
JCSCS - Journal of Computer Science and Control Systems, Vol. 8, Nr. 2, October 2015
JCSCS - Journal of Computer Science and Control Systems, Vol. 8, Nr. 2, October 2015
Academy of Romanian Scientists
C
S
University of Oradea, Faculty of Electrical Engineering and Information Technology Vol. 8, Nr. 2, October 2015
Journal of Computer Science and Control Systems
ISSN: 1844-6043
University of Oradea Publisher
University of Oradea Publisher
Academy of Romanian Scientists
C
S S
University of Oradea, Faculty of Electrical Engineering and Information Technology Vol. 8, Nr. 2, October 2015
Journal of Computer Science and Control Systems
University of Oradea Publisher
2 Volume 8, Number 2, October 2015 ___________________________________________________________________________________________________________
EDITOR IN-CHIEF Eugen GERGELY - University of Oradea, Romania EXECUTIVE EDITORS Gianina GABOR
- University of Oradea, Romania
Daniela E. POPESCU - University of Oradea, Romania
Helga SILAGHI
- University of Oradea, Romania
Viorica SPOIALA
- University of Oradea, Romania
ASSOCIATE EDITORS Mihail ABRUDEAN Angelica BACIVAROV Valentina BALAS Eugen BOBAŞU Dumitru Dan BURDESCU Toma Leonida DRAGOMIR János FODOR Voicu GROZA Štefan HUDÁK Geza HUSI Ferenc KALMAR Jan KOLLAR Mohamed Najeh LAKHOUA Anatolij MAHNITKO Ioan Z. MIHU Emilia PECHEANU Constantin POPESCU Dumitru POPESCU Luminiţa POPESCU Alin Dan POTORAC Ioan ROXIN Ioan SILEA Lacramioara STOICU-TIVADAR Lorand SZABO Janos SZTRIK Honoriu VĂLEAN
ISSN 1844 - 6043
Technical University of Cluj-Napoca, Romania University Politehnica of Bucharest, Romania "Aurel Vlaicu" University of Arad, Romania University of Craiova, Romania University of Craiova, Romania "Politehnica" University of Timisoara, Romania Szent Istvan University, Budapest, Hungary University of Ottawa, Canada Technical University of Kosice, Slovakia University of Debrecen, Hungary University of Debrecen, Hungary Technical University of Kosice, Slovakia University of Carthage, Tunisia Riga Technical University, Latvia "Lucian Blaga" University of Sibiu, Romania "Dunărea de Jos" University of Galaţi, Romania University of Oradea, Romania University Politehnica of Bucharest, Romania "Constantin Brâncuşi" University of Tg. Jiu, Romania "Stefan cel Mare" University of Suceava, Romania Universite de Franche-Comte, France "Politehnica" University of Timisoara, Romania "Politehnica" University of Timisoara, Romania Technical University of Cluj Napoca, Romania University of Debrecen, Hungary Technical University of Cluj-Napoca, Romania
This volume includes papers in the following topics: Bayesian networks, BrushLess Direct Current motors and drives, Churn prediction, Control engineering, Cyber-physical systems, Databases and information systems, Dependable computing, Digital control, Fault tolerant systems, Fuzzy reasoning, Fuzzy Petri nets, Parameters and structure learning, Reliability, Safety-critical systems, Server virtualization, Servo systems, Structured analysis, System modeling, simulation and control, Telecommunications.
Journal of Computer Science and Control Systems 3 ___________________________________________________________________________________________________________
CONTENTS
AL-FALAHY Raed I. Hamed - University of Human Development, Sulaymaniyah, Iraq Ethylene and Ethane Analysis Based on the Fuzzy Method to Identify Oil Thermal Criticality of the Transformer ..........5 BRANDUSOIU Ionut, TODEREAN Gavril - Technical University of Cluj-Napoca, Romania Churn Prediction in the Telecommunications Sector Using Bayesian Networks ............................................................11 DOBRA Mirela, AŞTILEAN Adina, ABRUDEAN Mihail - Technical University of Cluj-Napoca, Romania Model Based Control for Applications with BLDC PM Motors ........................................................................................19 GORDIN Ionel - University "Ştefan cel Mare" of Suceava, Romania Optimized P2V Conversion Using VMware Converter Standalone ................................................................................24 KOVENDI Zoltan1, GERGELY Eugen-Ioan1, HUSI Geza2, COROIU Laura1 - 1 University of Oradea, Romania, University of Debrecen, Hungary Algorithm for Modeling and Checking Differential Equations in Nonlinear Process Control ...........................................29
2
MBIHI Jean - University of Douala, Cameroon A Flexible Multimedia Workbench for Digital Control of Input-Delay Servo Systems .....................................................35 VARI-KAKAS Ştefan, POSZET Otto - University of Oradea, Romania Considerations Regarding the Design and Reliability Analysis of Safety Critical Systems ............................................41
4 Volume 8, Number 2, October 2015 ___________________________________________________________________________________________________________
Journal of Computer Science and Control Systems 5 ___________________________________________________________________________________________________________
Ethylene and Ethane Analysis Based on the Fuzzy Method to Identify Oil Thermal Criticality of the Transformer AL-FALAHY Raed I. Hamed University of Human Development, Sulaymaniyah, Iraq, Department of Computer Science, College of Science and Technology, Sulaymaniyah, Iraq, E-Mail:
[email protected]
Abstract - The failure mechanisms in oil immersed power transformers are critical factors. Ethylene and ethane analysis based on the Fuzzy Petri Net (FPN) approach has been developed for transformer assessment in transformer oil. In this paper we present a novel approach of FPN model to assess the transformer maximum practicable operating efficiency based on the values of ethylene and ethane of a rulebased system, using a fuzzy algorithm. The model is based on the theory of FPNs is used to describe a formal model of knowledge representation of uncertain and vague domains. The proposed approach is to bring ethylene and ethane analysis within the framework of a powerful modeling tool FPN. The two input features of our fuzzy model— the x(t) and u(t) at times instance t can be formulated as uncertain fuzzy tokens to determines the x(t+1) values at time instance (t+1). The FPN components and functions are mapped from the different type of fuzzy operators of If-parts and Thenparts in fuzzy rules. Issues relating to the application of the expression model to evolutionary computation are discussed. Keywords: oil thermal criticality; fuzzy reasoning; fuzzy Petri net; transformer. I. INTRODUCTION Power transformers are a vital link with a power system. Monitoring and diagnostic techniques are essential to decrease maintenance and improve reliability of the equipment [1]. Oil transformer is the key element in the transmission system for maintaining the power system reliability. The maintenance can only be based on monitoring with extended analytical and electrical tests that can define the service condition of the transformer and predict its further life expectancy. In the literature, several models of fuzzy inference were obtained for diagnosis of power transformer insulation has been studied by various researchers [2, 3, 4, 5]. Fuzzy model showing the relationship between the input and output variables is very useful in analysis, synthesis, and implementation of fuzzy diagnostic approach. FPN is a relatively new artificial intelligence technique that is applied for condition monitoring [2, 3]. FPN means approximate reasoning, information
granulation, computing with logical words and so on. FPN is a novel fuzzy-based approach that deals with heterogeneous data onto much linguistic and numeric types, imprecise, vague in-formation, and concepts encountered in the mechanical-fit process and facilitate the expression of the reasoning process of an experienced observer with minimal rules. Petri nets theory and fuzzy logic exhibit a graphical and mathematical formalism to model, and simulate the biological systems. FPN is a successful tool for describing and studying information systems. Incorporating the fuzzy logic in FPN has been widely used to deal with fuzzy knowledge representation and reasoning [6, 7]. It has also proved to be a powerful representation method for the reasoning of a rule-based system. An oil thermal criticality model is developed for testing the transformer to identify its oil thermal criticality as shown in Fig. 1. The explanation of how to reach conclusions is expressed through the movements of tokens in FPNs [7, 8]. The field of fuzzy Petri nets may have an important impact in understanding how transformer diagnostic system works, giving at the same time a way to describe, manipulate, and analyse them. Given the complexity of systems begin studied, engineers need a modeling and simulation framework to make sense of large-scale data and intelligently design traditional bench-top experiments that provide the most engineers insight. FPN as a new tool for predicting the oil thermal criticality values at time instance (t+1) for each input variables gases of ethylene (C2H4) x(t) and ethane (C2H6) u(t) at times instance t are investigated in this paper. In order to validate our approach, we compare our method to the fuzzy logic toolbox of MATLAB. The comparison is made in terms of the oil thermal criticality value measure of the input variables. The similarity that we have discovered is that they both have the same conclusions. The structure of this paper is as follows: In Section 2, fuzzy Petri nets are described. In Section 3, the formulation of fuzzy sets and linguistic variables are presented. In Section 4, we explain the details of methods of modeling transformer diagnostic system. Section 5 describes experimental and simulation results. Finally, we presented the conclusions of model in Section 6.
6 Volume 8, Number 2, October 2015 ___________________________________________________________________________________________________________
II. FUZZY PETRI NET MODEL A. Formal definition of fuzzy Petri nets Upon critically examining the past and current literature, one will notice that no one has studied the FPN aspects of diagnostic the transformer models so far. Therefore, in this paper, we address the FPN aspects [7] can have a good idea about the relative merits and demerits of each model of diagnostic. A FPN model allows a structural representation of knowledge and has got a systematic procedure for supporting the fuzzy reasoning process Formally, a fuzzy Petri net structure is defined as follows [7]: The tuple FPN = (P, T, D, I, O, F, α, β) is called a fuzzy Petri net if: 1. P = {pl, p2 , ..., pn} is a finite set of places, corresponding to the propositions of FPRs; 2. T = {tl, t2 , ..., tn} is a finite set of transitions, P ∩T = Ø , corresponding to the execution of FPRs; 3. D = {d1, d2 , ..., dn} is a finite set of propositions of FPRs. P ∩ T ∩D = Ø, |P | =| D |, di (i = 1,2,..., n) denotes the proposition that interprets fuzzy linguistic variables, such as: very low, low, lnorm , eug , hnorm ,as in our model; 4. I : P×T → {0, 1} is an n×m input incidence matrix defining the directed arcs from propositions (P) to rules (T). I(pi, tj) = 1, if there is a directed arc from pi to tj, and I(pi, tj) = 0; if there is no directed arcs from pi to tj, for i = 1, 2, …,n, and j = 1, 2,…,m. 5. O : P×T → {0, 1} is an n×m is an output incidence matrix defining the directed arcs from rules to propositions. O(pi, tj) = 1, if there is a directed arc from tj to pi, and O(pi, tj) = 0; if there is no directed arcs from tj to pi, for i = 1, 2, …,n, and j = 1, 2,…,m. 6. F = {µ1, µ2 ,..., µm} where µi denotes the certainty factor (CF =µi) of Ri , which indicates the reliability of the rule Ri , and µi ∈ [0,1]; 7. α : P → [0,1] is the function which assigns a token value between zero and one to each place; 8. β: P → D is an association function, a bijective mapping from a set of places to a set of propositions. Moreover, this model can be enhanced by including a function Th: T →[0, 1] which assigns a threshold value Th(tj) = λj ∈ [0, 1] to each transition tj, where j= 1,…, m. Further more, a transition is enabled and can be fired in FPN models when values of tokens in all input places of the transition are greater than its threshold [9, 10, 11, 12]. A token value in place pi ∈ P is denoted by α(pi) ∈ [0, 1], α(pi) = yi, yi ∈ [0, 1] and β(pi) = di. This states that the degree of the truth of proposition di is yi. A transition ti is enabled if ∀ pi ∈ I(ti), yi > 0. If this transition ti is fired, tokens are removed from input places I(ti) and a token is deposited onto each of the output places O(ti). This token’s membership value to the place pk, (i.e. yk = α(pk)), is part of the token and gets calculated within the transition function. It is easy to see that CF ∈ [0, 1]. In a fuzzy Petri net, different types of rules can be represented. The general ones are:
1) A simple fuzzy production rule: 2) IF di THEN dk (CFj = f (tj)); 3) A composite conjunctive rule: IF d1 ANDd2 AND … AND dj THEN dk (CFj = f (tj)); 4) A composite disjunctive rule: IF d1 OR d2 OR … OR dj THEN dk (CFj = f (tj)); B. Oil thermal criticality model The Mamdani fuzzy inference system [13] was proposed as the first attempt to control a steam engine and boiler combination by a set of linguistic control rules obtained from experienced human operators. The output membership functions of Mamdani models are fuzzy sets, which can incorporate linguistic information into the model. The computational approach described in this paper is Mamdani fuzzy Petri net (MFPN) that is able to overcome the drawbacks specific to pure Petri nets. Fuzzy models describing dynamic processes compute the states x(t+1), at a time instant t+1, from the information of the inputs x(t) and u(t), at time instant t as sown in Eq. (1), x(t + 1) = f (x(t), u(t)) (1) where f (·) is a fuzzy model with the structure shown in Fig. 1. Input Layer (Layer 1), as shown in Eq. (2), no calculation is done in this layer. Each node, which corresponds to the inputs, x(t) and u(t), only transmits input value to the next layer directly. The certainty factor of the transitions in this layer is unity.
O i(1) = x ( t ), u ( t )
(2)
where x(t) and u(t) are the value of the ith oil thermal (1)
criticality at time instant t, and Oi is the ith output of layer 1. The values of the inputs, x(t) and u(t), and of the outputs, x(t+1), can be assigned linguistic labels. The output link of layer 2, represented as the membership value, specifies the degree to which the input value belongs to the respective label. Linguistic rules can be formulated that connect the linguistic labels for x(t) and u(t) via an IF-part, called an antecedent of a rule and the THEN-part, also called a consequent of the rule which determines the resulting linguistic label for x(t+1). The antecedent membership functions are the membership functions appearing in the IF-part of the rule in layer 2 and the consequent membership functions are the membership functions appearing in the THENpart in layer 4. As shown in Eq. (3), 0, x ≤ a.
( x − a ) /(b − a ), a ≤ x ≤ b. (3) b≤x≤c µ A (x) = 1 (d − x ) /(d − c), c ≤ x ≤ d. 0, d ≤ x. where µA refers to the degree to which x belongs to the linguistic label A and the parameters {a, b, c, d} (with a corners of the underlying trapezoidal membership function. The node in layer 3 combines the antecedent part of a fuzzy rule using a T-norm operator. In this a
