Inter national J our nal of Electrical and Computer Engineering (IJECE) V ol. 6, No. 4, August 2016, pp. 1529 1533 ISSN: 2088-8708, DOI: 10.11591/ijece.v6i4.10704 1529       I ns t it u t e  o f  A d v a nce d  Eng ine e r i ng  a nd  S cie nce   w     w     w       i                       l       c       m     A Criterion on Existence and Uniqueness of Beha vior in Electric Cir cuit T akuya Hirata * , Ek o Setiawan * , Kazuya Y amaguchi ** , and Ichijo Hodaka *** * Interdisciplinary Graduate School of Agriculture and Engineering, Uni v ersity of Miyazaki ** Department of Control Engineering, National Institute of T echnology Nara Colle ge *** Department of En vironmental Robotics, F aculty of Engineering, Uni v ersity of Miyazaki Article Inf o Article history: Recei v ed Feb 5, 2016 Re vised May 22, 2016 Accepted Jun 8, 2016 K eyw ord: beha vior of electric circuit switching circuit circuit analysis ABSTRA CT Beha vior of electric circuits can be observ ed by solving circuit equations symbolically as well as numerically . In general, symbolic computation for circui ts with certain number of circuit elements needs much more time than numerical computation. It is reasonable to check the e xistence and uniqueness of the solution to circuit equations beforehand in order to a v oid computation for the case of no solution. Indeed, some circuits ha v e no solution; in that case, one should notice it and a v oid to w ait meaningless computation. This paper proposes a ne w theorem to check whether gi v en circuit equations ha v e a solution and their v oltages and currents of all circuit elements are uniquely determined or not. The theorem is suitable for de v eloping a computer algorithm and helps quick symbolic computation for electric circuits. Copyright c 2016 Institute of Advanced Engineering and Science . All rights r eserved. Corresponding A uthor: Ichijo Hodaka Department of En vironmental Robotics, F aculty of Engineering, Uni v ersity of Miyazaki 1-1, Gakuen Kibanadai Nishi, Miyazaki, 889-2192, Japan hijhodaka@cc.miyazaki-u.ac.jp 1. INTR ODUCTION An y simulation of beha vior of an actual circuit is based on modelling of the circuit with idealization. The modelling is a crucial step of analysis and design; it depicts the actual c ircuit as a circuit diagram, determines a w orking point, and linearizes characteristics of electric components in the actual circuit with a w orking frequenc y . Beha vior of the actual circuit is represented by a solution to circuit equations deri v ed from the circuit diagram and the fundamental la ws such as Kirchhof f s la ws, Ohm’ s la w , and the electric characteristics of inductor and capacitor . SPICE[1], a defacto standard circuit simulator , simulates and plots the beha vior by numerical calculation. That is, we there assume e xistence and uniqueness of beha vior . W e also assume them if we measure beha vior of an actual circuit. Hence, a modelling process should not sacrifice e xistence and uniqueness of beha vior; v oltages and currents at components in the actual circuit or in the simulation model should be single-v alued functions defined on all time. Some studies address circuit diagrams which ha v e lost e xistence and uniqueness of beha vior . F or e xam- ple, some of inductor currents and capacitor v oltages in the circuit diagram are not eligible for a member of state v ariables[2]-[5]. This can be translated into a problem to find a spanning tree in the circuit by graph theory[3]-[5]. Ho we v er , little is kno wn about a direct procedure to check a gi v en circuit diagram. In this paper , we propose a quick criterion to check whether a solution to linear circuit equations e xists and is uniquely determined or not. The criterion enables us to v alidate complicated circuits such as a switching circuit with computers, and thus dra w a pr oper circuit diagram - a diagram which has e xistence and uniqueness of beha vior - with help of computers. Notice that the criterion is writt en in a symbolic equation. That means our result w ould be a fundamental result to contrib ute the field of symbolic calculation[6][7] of electric circuits. J ournal Homepage: http://iaesjournal.com/online/inde x.php/IJECE       I ns t it u t e  o f  A d v a nce d  Eng ine e r i ng  a nd  S cie nce   w     w     w       i                       l       c       m     Evaluation Warning : The document was created with Spire.PDF for Python.
1530 ISSN: 2088-8708 2. LA WS OF ELECTRIC CIRCUITS AND CIRCUIT EQ U A TIONS F or a gi v en circuit diagram, circuit equations are defined by the fundamental la ws, Kirchhof f s v oltage la w(KVL), Kirchhof f s current la w(KCL), Ohm’ s la w , and the electric characteristics of inductor and capacitor , where all v ariables in the la ws and the characteristics are assumed to be single-v alued functions of time t . In general, we ha v e Aw + B y + C z = 0 (1) H ( x ( t ) x ( t 0 )) = Z t t 0 z 1 ( p ) dp (2) where, A; B ; C ; and H are constant matrices whose entries are determined by the la ws and the characteristics of electric circuit and resistance of resistors, x = i L v C ; u = v V i I ; w = u x ; y : the v ector of node v oltages ; z 1 = v L i C ; z = 2 6 6 6 6 6 6 6 6 4 i V v I v R i R z 1 v S i S 3 7 7 7 7 7 7 7 7 5 ; (3) and v V ; v I ; v C ; v L ; v R , and v S are v oltages of v oltage sources, current sources, capacitors, inductors, resistors and switches, and i V ; i I ; i C ; i L ; i R , and i S are currents of v oltage sources, current sources, capacitors, inductors , resistors and switches, and v N is potential of nodes, respecti v ely . This paper assumes that all the v ariables are Riemann inte grable, and thus the notation R in the equation (2) is Riemann inte gral (for definition of inte gral, see e.g. [8]). N 1 N 3 N 0 N 2 Figure 1. A boost con v erter - CCM ( S 1 : ON, S 2 : OFF) Example Figure 1 and 2 represent so-called continuous conduction mode (CCM) and discontinuous conduction mode (DCM) of a boost con v erter [9], respecti v ely . Circuit equations of the circuit diagram in Figure 1 are represented IJECE V ol. 6, No. 4, August 2016: 1529 1533 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 1531 in the form (1) and (2), where A = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 ; B = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 1 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 ; C = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 1 0 0 0 1 0 1 R 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 ; (4) H = L 0 0 C ; x = i L v C ; z 1 = v L i C ; w = 2 4 v V i L v C 3 5 ; y = 2 6 6 4 v N 1 v N 3 v N 0 v N 2 3 7 7 5 ; and z = 2 6 6 6 6 6 6 6 6 6 6 6 6 4 i V v R i R v L i C v S 1 i S 1 v S 2 i S 2 3 7 7 7 7 7 7 7 7 7 7 7 7 5 : (5) N 1 N 3 N 0 N 2 Figure 2. A boost con v erter - DCM ( S 1 : OFF , S 2 : OFF) Circuit equations of Figure 2 are represented in the form (1) and (2), where A = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 ; B = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 1 1 0 0 1 0 1 0 0 1 0 1 0 1 0 1 0 1 1 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 ; C = 2 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 4 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 0 0 1 0 1 0 1 0 1 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 1 0 0 0 1 0 1 R 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 3 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 7 5 (6) A Criterion on Existence and Uniqueness of Behavior in Electric Cir cuit (T akuya Hir ata) Evaluation Warning : The document was created with Spire.PDF for Python.
1532 ISSN: 2088-8708 and the other notations are already defined as (5). The v ariable v L cannot be uniquely determined on the algebraic circuit equations (6) of Figure 2 for an y v V ; i L , and v C , which is understood by looking at the circuit diagram. W e will propose a theorem to capture the abo v e situation by rank calculations of matrices A; B and C in the algebraic equation (1). 3. THE EXISTENCE AND UNIQ UENESS OF SOLUTION OF ALGEBRAIC CIRCUIT EQ U A TIONS If for an y w , a solution ( y ; z ) to t he algebraic circuit equation (1) of a circuit e xists and z is unique, then the circuit is said to be pr oper in this paper . A circuit is said to be impr oper if it is not proper . Theor em 1. A circuit is proper if and only if the coef ficient matrices of its algebraic equation in the form (1) satisfy rank [ A B C ] = rank [ B C ] = rank B + m: where m is the number of column of the matrix C . The proof of Theorem 1 is gi v en in the appendix A. Rank calcul ation in Theorem 1 is performed on general purpose symbolic computation system widely used. This helps us check whether a circuit is proper or not, e v en if it is lar ge and complicated, and then dif ficult to check by hand. Example (A pr oper cir cuit) W e recall the circuit in Figure 1 and obtain rank [ A B C ] = 12 ; rank [ B C ] = 12 ; rank B = 3 ; m = 9 for (4). Therefore the circuit is proper by Theorem 1. Example (An impr oper cir cuit) W e recall the circuit in Figure 2 and obtain rank [ A B C ] = 12 ; rank [ B C ] = 11 ; rank B = 3 ; m = 9 for (6). Therefore the circuit is improper by Theorem 1. The whole beha vior of a proper circuit is gi v en by Corollary 1 in the appendix A. 4. CONCLUSION Beha vior of an actual electric circ u i t, a set of v oltages and currents of all circuit components, is assumed to be e xpressed as a single-v alued function defined on all time, in general. Designers depict the actual circuit as a schematic diagram to simulate beha vior of the actual circuit. Careful designers model the circuit as a diagram with a decision of including parasitic elements or not and reflect a w orking point and a w orking frequenc y of the circuit on the diagram. The y write circuit equations equi v alent to the diagram, and then, solv e the equations to simulate beha vior of the actual circuit. Ho we v er , e xistence and uniqueness which are assumed to beha vior of the actual circuit are not necessarily inherited to the solution because the diagram and the equations are a reduction of the actual circuit. This paper has proposed a criterion for the e xistence and uniqueness to be guaranteed. The proposed criterion is e xpressed as equality between ranks of coef ficient matrices in the circuit equations. Ef fecti v e and quick calculation of matrix rank is a v ailable in general purpose symbolic computing tools. Therefore our result contrib utes to the field of symbolic computation of electric circuits. A LEMMA, COR OLLAR Y AND PR OOF Cor ollary 1. If a circuit is proper , the equations (1) and (2) are uniquely solv ed as: z = K 1 x + K 2 u (7) x ( t ) = e F ( t t 0 ) x ( t 0 ) + e F t Z t t 0 e F p Gu ( p ) dp: (8) Pr oof 1. W e immediately obtain the e xpression (7) by Theorem 1. This includes z 1 = K 11 x + K 21 u . Notice that the matrix H is block-diagonal whose blocks ha v e capacitances C s and self or mutual inductances ( L s and M s respecti v ely) in their entries and is in v ertible. If we put F = H 1 K 11 and G = H 1 K 21 and apply inte gration by parts[8] with a matrix e xponential function, we ha v e (8). W e remark that the solution (8) is obtained without assuming x ( t ) to be dif ferentiable. IJECE V ol. 6, No. 4, August 2016: 1529 1533 Evaluation Warning : The document was created with Spire.PDF for Python.
IJECE ISSN: 2088-8708 1533 Lemma 1. Let B and C be matrices wi th a common number of ro ws. The equation B y + C z = 0 has a solution z for an y y if and only if rank [ B C ] = rank C . Pr oof 2. (only if) There is a matrix Z such that B = C Z . Then [ B C ] = [ C Z C ] = C [ Z U m ] , where U m is an identity matrix. rank [ B C ] = rank C , because [ Z U m ] is ro w full rank. (if) Since rank [ B C ] = rank C and Im [ B C ] Im C , Im [ B C ] = Im C . Hence Im B Im [ B C ] = Im C . This means that for e v ery fix ed y , there is a z such that B y + C z = 0 . Pr oof 3 (F or Theorem 1) . (only if) By the assumption and Lemma 1, rank [ A B C ] = rank [ B C ] . Let w = 0 . The solution space Q of the equation (1) is Q = k er[ B C ] . Let y 1 z 1 2 Q and y 2 2 k er B . Since y 2 0 2 Q , z 1 = 0 . Hence, dim(k er[ B C ]) = dim(k er B ) . By rank-nullity theorem, we obtain rank [ B C ] = rank B + m: (if) Let Q ( w ) = f y z j Aw + B y + C z = 0 g . By Lemma 1, for e v ery fix ed w , Q ( w ) 6 = . Since rank [ B C ] rank B + rank C and rank C m in general, rank [ B C ] = rank B + m follo ws m = rank [ B C ] rank B rank C m . So, rank C = m = rank [ B C ] rank B : (9) Use a general equality rank [ B C ] = rank B + rank C dim( Im B \ Im C ) [10] and (9) sho w dim( Im B \ Im C ) = 0 : (10) Let y 1 z 1 and y 2 z 2 2 Q ( w ) . Then B y 1 + C z 1 = B y 2 + C z 2 . By (10), C ( z 1 z 2 ) = 0 . The first equality in (9) implies z 1 = z 2 . REFERENCES [1] L. W . Nagel and D. O. Pederson, SPICE (Simulation Program with Inte grated Ci rcuit Emphasis) , Memorandum No. ERLM382 Electronic Research Laboratory , 1973. [2] T . Bashk o w , “The A Matrix, Ne w Netw ork Description, IRE T ransactions on Circuit Theory , V ol. 4, No. 3, pp. 117-119, 1957. [3] E. S. K uh and R.A. Rohrer , “The State-V ariable Approach to Netw ork Analysis, Proceedings of the IEEE , V ol. 53, No. 7, pp. 672-686, 1965. [4] R. A. Rohrer , Circuit Theory: An Introduction to the State V ariable Approach , McGra w-Hill, 1972. [5] D. A. Calahan, Computer -Aided Netw ork Design , Re vised Edition, McGra w-Hill, 1972. [6] A. Luchetta, S. Manetti, and A. Reatti, “SAPWIN-a symbolic simulator as a support in electrical engineering education, IEEE T ransactions on Education , V ol. 44, No. 2, 2001. [7] T . Hirata, K. Y amaguchi, and I. Hodaka, A Symbolic Equation Modeler for Electric Circuits, A CM Communi- cations in Computer Algebra , V ol. 49, No. 3, Issue 193, 2015. [8] K. Knopp, Theory and Application of Infinite Series , Blackie, 1951. [9] R. W . Erickson and D. Maksimo vi ´ c, Fundamentals of Po wer Electronics , Second Edition, Kluwer , 2004. [10] P . Lancaster and M. T ismenetsk y , The Theory of Matrices , Second Edition, Academic Press, 1985. A Criterion on Existence and Uniqueness of Behavior in Electric Cir cuit (T akuya Hir ata) Evaluation Warning : The document was created with Spire.PDF for Python.