Mathematical Modelling of Engineering Problems

Vol. 7, No. 1, March, 2020, pp. 26-30

Journal homepage: http://iieta.org/journals/mmep

Circuit Modelling by Difference Equation: Pedagogical Advantages and Perspectives

Mohammad Reza Salehizadeh1*, Hassan Nouri2
1
Department of Electrical Engineering, Marvdasht Branch, Islamic Azad University, Marvdasht, Iran
2
Power Systems, Electronics and Control Research Laboratory, UWE Bristol BS16 1QY, UK

Corresponding Author Email: [email protected]

https://doi.org/10.18280/mmep.070104 ABSTRACT

Received: 13 October 2019 Circuit theory is a cornerstone course in electrical engineering and control majors in
Accepted: 26 January 2020 ordinary universities and colleges throughout the world. This course covers
fundamental principles and analysis methods of basic circuits commonly employed in
Keywords: the forthcoming courses. In most electrical programs after the introduction of basic
circuit, modelling, difference equation, dynamic elements of Ohm’s and Kirchhoff’s current and voltage laws, the dynamic response of
response, non-linear circuits the circuits containing capacitors and inductors will be studied. Customarily to solve
these circuits, advanced mathematical approaches such as differential equations are
used. Under such circumstances, the students are faced with two challenges, solving the
differential equations, and understanding the dynamic response of circuits. In order to
improve students’ understanding, an analysis tool with less mathematical prerequisites
should be used for the solutions before embarking on the use of conventional differential
equation techniques such as Laplace transform. Hence, we propose a novel approach
for these circuit analyses through the application of a discretized version of differential
equations which is used in discrete control systems. Although this approach has a well-
established background, its exploration uses in the circuit theory course as yet has not
been reported. The novelty of the proposed approach not only lies in its intuitive
simplicity but also in its contribution to the understanding and visualization of students
in the real-time response of linear and non-linear circuits to any desirable input without
any mathematical burden. The analysis can be performed by hand or this is also helpful
for those who prefer modern education aided by computers. This, in turn, may attract
more students to the program. In this paper, the effectiveness of the proposed approach
is demonstrated through a set of illustrative examples.

1. INTRODUCTION difference equation in the content of circuit analysis course.

In circuit theory course, after the introduction of R, L, C
According to the U.S. Bureau of Labor Statistics, combination, ohm’s law, and Thevenin’s theorem, the students
Employment Projections program, the employment of are familiarized with the analysis of first and second order
electrical engineers is projected to grow 7 percent from 2016 transient response of a circuit [10]. In this regard, the basic
to 2026 [1]. Taking any step towards improvement of the definitions, concepts, theorems, and tools for analysis of a
content and teaching style of the fundamental courses in circuit are instructed. To this purpose, acquiring fundamental
electrical engineering programs will enhance the scientific knowledge of physics, including the basics of electricity and
background of these new generation of engineers. Without magnetism, and mathematics, including knowledge of solving
doubt, circuit theory is one of the most important and algebraic and differential equations, are the prerequisite for
fundamental courses in the electrical engineering curriculum this course. Most of the students who participate in this course
throughout the world [2, 3]. Rather than electrical engineering have knowledge of basic algebraic equations from high school.
major, learning and research in a few other majors such as However, the students usually become familiar with
automation, and computer science is based on circuit theory differential equations through university courses. Having the
[4]. Specifically, in electrical engineering, the knowledge knowledge to solve differential equations is necessary in some
acquired in this course is a prerequisite for most of the parts of the circuit theory course such as analysis of RLC
advanced courses such as electronics, power system analysis, circuits. According to the information gathered from Iranian
and electrical machines. Given this introduction about the and some European instructors, in spite of having a
importance of circuit analysis course, many researches have background in differential equations, most students who take
been devoted to the success of the students in this course [5]. a circuit theory course do not have sufficient mastery in
These approaches include, but not limited to, improving solving these equations. Under such cases, instructing
teaching style [6, 7], text book selection [8], and improving dynamic analysis of the circuits faces some challenges. Due to
teaching materials [9]. Despite of these researches, further the lack of essential mastery on differential equation solution
steps need to be taken to fulfill the gaps in the content of this methods, the students are not able to fully focus on the analysis
important course. This paper intends to address this important of such circuits. This in turn may lead to decreased motivation
issue by proposing incorporation of circuit analysis through and attraction to this course. To overcome these challenges,

26
regardless of the students’ competence to solve differential especially when we face non-linearity. This approach has been
equations, a soft start topic with less mathematical used extensively for solving engineering problems. For
requirements needs to be introduced as an extension to this example, to design an optimal control method by means of a
course before embarking on solving directly the differential dynamic programming approach, the differential equations
equation. With this aim, we propose to initially familiarize need to be approximated by difference equations [11]. An
students with the conversion method of differential equations application for computer control of practical continuous-time
to difference equations by a discretization process. Thereafter, processes is proposed by Yang and Ding [12]. Dynamic
to analyze the obtained difference equations either by hand or behaviors of nonlinear models arising in ocean engineering are
by writing simple codes. In general, the proposed has a number treated through a similar way [13]. As shown in these papers
of benefits including increased appeal of circuit theory to [5, 6] in the engineering practice approximation by difference
students’ and provision of more time for students to focus on equation is used for solving non-linear and complicated
the concept and engineering analysis of the circuit without systems. However, as stated in the introduction, from an
facing mathematical challenges at the beginning of the course. educational perspective, not only for the analysis of nonlinear
This enhances their competence in the solution of differential circuits, but also for linear circuits, establishing an additional
equations, understanding real-time response of an electrical course based on difference equation has some merits.
circuit to any desirable input, and finally enabling them to
produce simple codes to understand non-linear circuits 2.2 Step-by-step procedure
without any mathematical burden. Although the modeling and
analysis of dynamic systems through difference equations has The proposed procedure of the educational course is as
a very established background, from an educational follows:
perspective the proposed approach is novel. To date no report
can be found in published educational literature where i) Instruction of modeling dynamic circuits: At this step by
differential equations represent the circuit model transformed using KVL and KCL laws, the set of differential equations is
into the difference equations in circuit theory course. derived.
The remainder of the paper is organized as follows; section ii) Instruction of discretization: The obtained differential
II discusses in brief the review on difference equations. A set equations from the previous step is discretized in the following
of illustrative examples are solved in section III where also the form:
advantages of the proposed approach is demonstrated. Section
IV concludes the paper. 𝜃(𝑘 + 1) = 𝐴𝜃(𝑘) + 𝐵𝑢(𝑘)⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡⁡𝑘 = 0,1, … , 𝑁 (1)

where, 𝜃 is the state vector that could be voltage or current and

2. THE PROPOSED MODELLING APPROACH 𝑢 is the input of the circuit. It is noticed that 𝜃 are 𝑛 × 1 vector
in which n is the order of the circuit.
2.1 Difference equation
iii) Instruction of what happens to the instantaneous
In simple words, a mathematical equation with a function voltages and currents which have been discretized and shown
and its related derivatives is called a “differential equation”. A as the state vector ( 𝜃 ). With this aim, the following
differential equation is usually used to model continuous pseudocode could be used to solve the recursive equation of
dynamic systems. In contrary, difference equation consists of (1).
the differences between successive values of a function of a
discrete variable. Two well-known mathematical transforms, Begin
Laplace and z-transform, are used to find the close-form
solution for differential and difference equations, respectively. Input A, B, 𝜃(0)
Solving a differential equation yields the analog signal, while
solving difference equations yields digital signals. Difference Do
equations are used to model discrete dynamic control systems
or to approximate continuous dynamic systems. Repeat 𝑘 = 0: 𝑁
Approximation of differential equations by difference
equations is useful for cases in which a close-form solution is 𝜃(𝑘 + 1) = 𝐴𝜃(𝑘) + 𝐵𝑢(𝑘)
not possible to be obtained for differential equations. For an
instant, consider a dynamic system that is presented with the End
following mathematical model: 𝑦 ′ (𝑡) + 𝑦(𝑡) = sin⁡(𝑡). From
𝑦(𝑡+∆𝑡)−𝑦(𝑡)
the definition of the derivative, we have 𝑦 ′ (𝑡) ≅ iv) Asking the students to develop codes for some of the
∆𝑡
where ∆𝑡 should be chosen as small as possible. By defining relevant exercises to obtain a better level of understanding.
𝑡 = 𝑘∆𝑡 in which 𝑘 = 0,1,2, … , we have 𝑦(𝑡) = 𝑦(𝑘) ,
𝑦(𝑘+1)−𝑦(𝑘)
𝑦 ′ (𝑡) ≅ , and sin(𝑡) = sin⁡(𝑘) . The equivalent 3. THE ILLUSTRATIVE EXAMPLES
∆𝑡
𝑦(𝑘+1)−𝑦(𝑘)
difference equation would be + 𝑦(𝑘) =
∆𝑡
In this section, we will provide three simple educational
sin(𝑘) , 𝑘 = 0, 1,2, …. The smaller value for ∆𝑡 results in a
examples to show the advantages of the proposed approach.
more precise approximation. This approximation can be
employed for both linear and non-linear differential equations.
3.1 Linear-time invariant circuit
Given the initial condition, the resultant difference equation
can be implemented through a computer code. This approach
The schematic of RL circuit is shown in Figure 1.
for solving differential equations becomes important,

27
 The same result can be obtained through direct solution of
conventional differential equations [3].

Figure 1. A simple RL circuit

Following the step-by-step procedure of II-B, the dynamic

equation for this circuit is formed by Kirchhoff’s voltage law
(KVL):

𝑑𝑖(𝑡)
𝑅𝑖(𝑡) + 𝐿 = 𝑉(𝑡) (2)
𝑑𝑡 Figure 2. Current of a simple RL circuit resultant from
difference equation
The differential equation is transformed to a difference
equation. Dividing the time interval [0, t] into N increments of
3.2 Non-linear-time invariant circuit
∆𝑡 , yields 𝑡 = 𝑘∆𝑡 , in which 𝑘 = 0,1, … , 𝑁 . The resultant
difference equation becomes:
To show the effectiveness of the proposed methodology for
𝑖(𝑘 + 1) − 𝑖(𝑘)
instruction of non-linear circuits, assume that the inductance
𝑅𝑖(𝑘) + 𝐿 = 𝑉(𝑘) (3) of the circuits shown in Figure 1 is now nonlinear that has the
∆𝑡
(k=0,1,…,N) flux linkage of:

100𝜆3
The re-arrangement of (3) is: i(𝜆) = 95𝜆5 + + 𝜆.
𝜆+5

𝑅∆𝑡 ∆𝑡
𝑖(𝑘 + 1) = (1 − ) 𝑖(𝑘) + 𝑉(𝑘) (4) Applying KVL results the following nonlinear differential
𝐿 𝐿
equation:
Considering 𝑅 = 1Ω, 𝐿 = 0.1⁡𝐻, ∆𝑡 = 0.001⁡𝑠𝑒𝑐𝑜𝑛𝑑, 𝑡 =
𝑑𝜆 100𝜆3
1⁡𝑠𝑒𝑐𝑜𝑛𝑑, and the input is a step function (𝑉(𝑡) = 𝑢(𝑡)), Eq. + 95𝜆5 + + 𝜆 = 𝑢(𝑡) (5)
(4) yields as 𝑖(𝑘 + 1) = 0.98𝑖(𝑘) + 0.01𝑢(𝑘) . This 𝑑𝑡 𝜆+5
difference equation is solved using the pseudocode of section
II. The students can also solve this equation by hand. At t =0 This type of example is not expected to be solved by
second, considering⁡𝑘 = 0, 𝑢(0) = 1 and the initial current students who are taking this course. However, here it is solely
𝑖(0) = 0, then 𝑖(1) = 0.01. Similarly at t=1 second, 𝑘 = 1, used as an educational example. Analysis of such an example
𝑢(1) = 1 and 𝑖(1) = 0.01, then 𝑖(2) = 0.0198, and so on. could easily be understood via the proposed procedure in
which the discretization of Eq. (5) results in the following
As it is observed the students are able to easily follow what is
difference equation:
happening in real-time simulation and how the instantaneous
values of the input voltage and currents affect the result. Such
0.1𝜆3 (𝑘)
observation and analysis are not easily possible via direct 𝜆(𝑘 + 1) = −0.095𝜆5 (𝑘) + 0.999𝜆(𝑘) −
𝜆(𝑘) + 5 (6)
solution of the differential equation. Additionally, the students
+ 0.001𝑢(𝑘)
can easily code this procedure in any programming device or
software. As an example the following code is written for
The solution for the Eq. (6) is similar to the earlier example
MATLAB environment that is related to the problem of Figure
in which at t= 0 second, 𝑘 = 0, 𝑢(0) = 1 and 𝜆(0) = 0, then
1. Its corresponding result is shown in Figure 2.
𝜆(1) = 0.001 . Also at t=1 second, 𝑘 = 1 , 𝑢(1) = 1 and
clear all; clc; close all 𝜆(1) = 0.001, then 𝜆(2) = 0.002 as shown below: 𝜆(2) =
0.1(0.001)3
−0.095(0.001)5 + 0.999(0.001) + + 0.001 =
R=1; L=0.1; (0.001)+5

dt=1e-3; 0.002. and so on. The MATLAB code for this problem is
illustrated below.
t=0:dt:1;
y=[0]; clear all;clc; close all
for k=1:length(t)-1 dt=1e-3;
y(k+1)=(1-R*dt/L)*y(k)+dt*1/L; t=0:dt:1;
end lambda=[0];
plot(t,y,'*') i=[0];
for k=1:length(t)-1

28
 lambda(k+1)=-.095*lambda(k)^5+0.999*lambda(k)-.1*la ∆𝑡
𝑖 (𝑘 + 1) 1
mbda(k)^3/(lambda(k)+5)+.001; [𝐿 ]=[ 𝐿 ] [𝑖𝐿 (𝑘)] (8)
𝑉𝑐 (𝑘 + 1) −∆𝑡 ∆𝑡 𝑉𝑐 (𝑘)
i(k+1)=95*lambda(k+1)^5+100*lambda(k)^3/(lambda(k) 1−
𝐶 𝑅𝐶
+5)+lambda(k);
end By setting the initial condition and parameters as 𝑖𝐿 (0) =
plot(t,lambda,'*') 0, 𝑉𝐶 (0) = 0, 𝑅 = 1⁡Ω, 𝐿 = 0.1⁡𝐻, 𝐶 = 0.001⁡𝐹, ∆𝑡 =
ylabel('\lambda(Wb)') 0.001⁡𝑠𝑒𝑐𝑜𝑛𝑑⁡𝑎𝑛𝑑⁡𝑡 = 1⁡𝑠𝑒𝑐𝑜𝑛𝑑 . The inductance current,
xlabel('Time(sec)') capacitor voltage, and the state trajectory are obtained and
plot(t,i,'*') shown in Figures 4-6, respectively.
ylabel('i(A)')
xlabel('Time(sec)')

Figures 3 and 4 show the magnetic flux of the nonlinear

inductance and current of the circuit, respectively.

Figure 5. Current of the inductor in RLC circuit

Figure 3. RLC circuit without external resources

Figure 6. State trajectory of RLC circuit

Figure 4. Current of the inductor in RLC circuit 4. CONCLUSION

3.3 Second order R-L-C circuit In this paper, we proposed circuit analysis through
difference equations which is frequently used in discrete
The proposed educational approach is able to be applied to control approaches. The novelty and effectiveness of the
higher order circuits. To show this capability, an RLC circuit proposed approach have been demonstrated. The approach is
without external resources is considered as shown in Figure 3. simple and complementary for the teaching of a typical circuit
By applying KCL the following state-space equations will be analysis course before embarking on the use of differential
derived for this circuit: equation approach. Furthermore, the proposed approach has a
number of benefits including increased student attraction to
𝑑𝑖𝐿 1 circuit theory and provision of more time for students to focus
0
[ 𝑑𝑡 ] = [ 𝐿 ] [ 𝑖𝐿 ] (7) on the concept and engineering analysis of circuits without
𝑑𝑉𝑐 −1 −1 𝑉𝑐
facing mathematical challenges at the earlier stage of the
𝑑𝑡 𝐶 𝑅𝐶 course. This will enhance their competence in the solution of
differential equations, their understanding of real-time
Discretizing these equations yields
response of an electrical circuit to any desirable input, and will
enable them to produce simple codes to understand linear and

29
non-linear circuits without any mathematical burden such as blended learning approach to course design and
Laplace Transform. Also, the approach could be applied to any implementation. IEEE Transactions on Education, 52(1):
circuit with any order. Moreover, as shown in the developed 19-30.
codes, the used commands are very basic and simple and a [7] Liu, Z., Zhao, W.Y., Xie, H., Zhong, H.S., Cui, H.L.
student who has passed a basic programming course will be (2018). Research of a multi-element teaching method in
able to write such codes by any programming package without retaking course of circuit analysis. Journal of Electrical
the need for in-depth mathematical knowledge. The developed & Electronic Education, 2: 31.
educational course could be applied in other engineering [8] Sangam, D., Jesiek, B K. (2014). Conceptual gaps in
disciplines for which the dynamics of the systems need to be circuits textbooks: A comparative study. IEEE
instructed. For instance, vibration theory course in mechanical Transactions on Education, 58(3): 194-202.
engineering, and introduction to heat transfer course in https://doi.org/10.1109/TE.2014.2358575
chemical engineering. [9] Johnson, A.M., Butcher, K.R., Ozogul, G., Reisslein, M.
(2013). Introductory circuit analysis learning from
abstract and contextualized circuit representations:
REFERENCES Effects of diagram labels. IEEE Transactions on
Education, 57(3): 160-168.
[1] Bureau of Labor Statistics, U.S. Department of Labor. https://doi.org/10.1109/TE.2013.2284258
(2020). Occupational Outlook Handbook, Electrical and [10] Parent, D.W. (2017). Novel gateway stay/add policy
Electronics Engineers. used to increase student success rates in an introductory
https://www.bls.gov/ooh/architecture-and- circuits class. In 2017 IEEE Frontiers in Education
engineering/electrical-and-electronics-engineers.htm, Conference (FIE), Indianapolis, IN, USA, pp. 1-8.
accessed on October 12, 2019. https://doi.org/10.1109/FIE.2017.8190600
[2] Kirk, D.E. (2004). Optimal Control Theory: An [11] Ge, Y., Li, S., Chang, P. (2018). An approximate
Introduction. Courier Corporation. dynamic programming method for the optimal control of
[3] Desoer, C.A. (2010). Basic Circuit Theory. Tata Alkai-Surfactant-Polymer flooding. Journal of Process
McGraw-Hill Education. Control, 64: 15-26.
[4] Xu, J.Q., Li, K.N., Liang, Q.F. (2018). Practice and https://doi.org/10.1016/j.jprocont.2018.01.010
research on the reform of teaching methods in circuit [12] Yang, D., Ding, R. (2014). Transforms from differential
course. In 2018 International Seminar on Education equations to difference equations and vice-versa applied
Research and Social Science (ISERSS 2018), Atlantis to computer control systems. Applied Mathematics
Press. https://doi.org/10.2991/iserss-18.2018.38 Letters, 31: 18-24.
[5] Parent, D.W. (2018). Examination the impact of various https://doi.org/10.1016/j.aml.2013.12.011
factors on student success in an introduction to circuit [13] Sulaiman, T., Bulut, H., Yokus, A., Baskonus, H. (2019).
analysis course. In 2018 IEEE Frontiers in Education On the exact and numerical solutions to the coupled
Conference (FIE), San Jose, CA, USA, pp. 1-5. Boussinesq equation arising in ocean engineering. Indian
https://doi.org/10.1109/FIE.2018.8659259 Journal of Physics, 93(5): 647-656.
[6] Hoic-Bozic, N., Mornar, V., Boticki, I. (2008). A https://doi.org/10.1007/s12648-018-1322-1

30