Submission 2 Joint Project
Non-parametric identification
of the DC motor
ACADEMIC Last
Mechatronics Engineering 19/02/2025
PROGRAM Update
COURSE
COURSE Digital Control PREAR2401MC01
CODE
NAME OF Parametric identification LABORATORY
1
LABORATORY of DC Motor NUMBER
PROFESSOR Mario Giraldo SEMESTER 2025-1
PREREQUISITE Analog Control DATE 14/03/2025
CO REQUISITE N/A HOURS PER/WEEK 2
1. Objectives
To obtain the mathematical model representing the dynamic behavior of a DC motor, with
an emphasis on speed dynamics.
To characterize the DC motor through the identification of a non-parametric model, which
in turn strengthens the understanding of the relationship between input and output signals
and the operation of the dynamic system.
To apply non-parametric techniques to one of the most widely used electrical devices in
academia and industry.
2. Materials and Tools
Computer with Labview, Matlab, and Simulink software, oscilloscope, regulated DC source, mul-
timeter, DC Motor integrated with encoder, L298N driver, myRio data acquisition card, LM331
integrated circuit.
The motor datasheet can be downloaded at the following link:
https://www.pololu.com/file/0J1736/pololu-37d-metal-gearmotors-rev-1-2.pdf.
For more details about the specific reference assigned, you can explore the following website:
https://www.pololu.com/product/4751/specs#note3.
3. Procedure
The process of identifying a DC motor involves determining the mathematical model of the dy-
namic system based on measurements of input and output, voltage, and speed in this case. The
1
Submission 2 Joint Project
Non-parametric identification
of the DC motor
mathematical model to be obtained is an approximate representation of the system’s real dynamics
with which its operation can be simulated and validated.
3.1. Non-Parametric Identification (Graphical Method)
Non-parametric identification methods are characterized because the models are constructed from
curves and their characteristics, illustrating the dynamic behavior of a system in response to a
determined input signal. These methods include transient analysis, frequency analysis, correlation
analysis, and spectral analysis.
In the development of this practice, the reaction curve method (transient analysis) will be applied,
and the model corresponding to the system’s response to a step input will be used as the model.
Identification will be performed using the approximation to a First Order plus Delay Transfer
Function (FOPDT) and a Second Order plus Delay Transfer Function (FOPDT), in Spanish known
as POR and SOR as seen in class.
3.1.1. First Order Plant with Delay (POR Model)
4. Objectives
To obtain the mathematical model representing the dynamic behavior of a DC motor, with
an emphasis on speed dynamics.
To characterize the DC motor through the identification of a non-parametric model, which
in turn strengthens the understanding of the relationship between input and output signals
and the operation of the dynamic system.
To apply non-parametric techniques to one of the most widely used electrical devices in
academia and industry.
5. Materials and Tools
Computer with Labview, Matlab, and Simulink software, oscilloscope, regulated DC source, mul-
timeter, DC Motor integrated with encoder, L298N driver, myRio data acquisition card, LM331
integrated circuit.
The motor datasheet can be downloaded at the following link: https://www.pololu.com/file/
0J1736/pololu-37d-metal-gearmotors-rev-1-2.pdf.
For more details about the specific reference assigned, you can explore the following website:
https://www.pololu.com/product/4751/specs#note3.
2
Submission 2 Joint Project
Non-parametric identification
of the DC motor
6. Procedure
The process of identifying a DC motor involves determining the mathematical model of the dy-
namic system based on measurements of input and output, voltage, and speed in this case. The
mathematical model to be obtained is an approximate representation of the system’s real dynamics
with which its operation can be simulated and validated.
6.1. Non-Parametric Identification (Graphical Method)
Non-parametric identification methods are characterized because the models are constructed from
curves and their characteristics, illustrating the dynamic behavior of a system in response to a
determined input signal. These methods include transient analysis, frequency analysis, correlation
analysis, and spectral analysis.
In the development of this practice, the reaction curve method (transient analysis) will be applied,
and the model corresponding to the system’s response to a step input will be used as the model.
Identification will be performed using the approximation to a First Order plus Delay Transfer
Function (POR) and a Second Order plus Delay Transfer Function (SOR).
6.1.1. First Order Plant with Delay (POR Model)
The transfer function corresponding to this type of plant is given by:
Y (s) Ke−θs
G(s) = =
U (s) τs + 1
Where K is the plant gain, τ is the system time constant, θ is the delay or dead time, y(t) is the
system output, and u(t) is the system input, with Y (s) and U (s) being the Laplace transforms
respectively.
The experimental procedure to estimate the model consists of opening the control loop (setting the
controller to manual if the system has one) before the final control element and creating a small
and quick step change in the process. The system response is plotted, and from the obtained cur-
ve, analysis is performed to estimate the values of the gain, the time constant, and the process delay.
In this case, the DC motor is not part of a control system, so there is no control loop to open;
therefore, it suffices to apply a step as input and obtain its speed response. Both signals should be
plotted.
On the curves obtained as a response, two representative points are located where the response
reaches 28.3 % and 63.2 % of its final value (See figure 1). These points are located at the values
corresponding to the elapsed times from the moment of the step application to the final
3
Submission 2 Joint Project
Non-parametric identification
of the DC motor
control element or to the process directly, and meet the following relations:
τ
θ+ = t1
3
θ + τ = t2
Figura 1: Graph of the input and output of the first-order system.
The values of t1 and t2 are calculated directly from the graphs or from the obtained database.
By solving equations 1 and 2 simultaneously, the values of τ and θ are estimated. The gain K is
obtained from the following expression:
∆Y
K=
∆U
6.1.2. Second Order Plant with Delay (SOR Model)
he transfer function corresponding to this type of plant varies depending on the value of the
damping constant ξ, and one of the following cases is presented:
Y (s) wn2 Ke−θs
G(s) = = 2 ξ<1
U (s) s + 2ξwn s + wn2
4
Submission 2 Joint Project
Non-parametric identification
of the DC motor
Y (s) Ke−θs
G(s) = = ξ≥1
U (s) (τ1 s + 1)(τ2 s + 1)
√
ξ± ξ 2 −1
Where τ1 , τ2 = wn
and K = ∆Y
∆U
.
For this case, three representative points must be located where the response reaches 15 %, 45 %,
and 75 % of its final value (See figure 2). These points are presented in the elapsed times from
the moment of the step application to the final control element or to the process directly.
Figura 2: Graph of the input and output of the second-order system.
t1 , t2 , and t3 are located similarly to the procedure done in the first-order system and calculate:
t2 − t1
x=
t3 − t1
With this value, and with ∆Y and ∆U , the other parameters of the second-order transfer function
are found from the following expressions:
0,0805 − 5,547(0,475 − x)2
ξ=
x − 0,356
F2 (ξ) = 2,6ξ − 0,6 ξ ≥ 1
or
5
Submission 2 Joint Project
Non-parametric identification
of the DC motor
F2 (ξ) = 0,708(2,811)ξ − 0,6 ξ < 1
F2 (ξ)
wn =
t2 − t1
F3 (ξ) = 0,922(1,66)ξ
F3 (ξ)
θ = t2 −
wn
6.1.3. Signal Conditioning of the Motor for myRio Card Use
Using the myRio acquisition card and Labview software, the dataset of the input voltage and output
speed measurements of the motor can be graphed and saved. For this, the following considerations
must be taken into account and the following instructions followed:
Input signal to the DC motor
The input signal to the DC motor for non-parametric identification will be a PWM with an am-
plitude of 5 volts that must be programmed on the myRio card. For this, it is necessary to consult
which pins are enabled for this purpose.
This PWM signal must be sent to the L298N driver, which must be powered with 12 VDC. The DC
motor must be connected to one of the driver’s motor outputs and must be configured appropriately.
A Labview interface must be programmed to apply steps that vary the percentage of the PWM
signal’s duty cycle. So, the input scale will be from 0 % to 100 %, which is suitable for the non-
parametric identification method.
Output signal of the DC motor
The encoder channels deliver a signal in the form of a pulse train whose frequency varies according
to the rotation speed of the motor, so it is necessary to condition the signal in such a way that a
reaction curve corresponding to an electrical signal proportional to the rotation frequency of the
motor and therefore to its speed is obtained.
The encoder module is configured to read the channels of the DC motor encoder. The encoder
of the myRio card delivers data corresponding to an incremental count of the encoder pulses, so
an adjustment must be made to calculate the difference between the current pulse value and the
previous one so that at each sampling instant the speed of that instant can be calculated. Once
6
Submission 2 Joint Project
Non-parametric identification
of the DC motor
this adjustment is made, the data called Encoder Pulses P E is obtained and the following general
formula for converting frequency to speed is applied:
PE 60s
× 1min
v= PPR
Ts
Where P E are the encoder pulses, P P R are the pulses per revolution, and Ts is the sampling
period in seconds to which the myRio has been configured. It should be noted that for different
programmable devices such as Arduinos, Raspberries, PICs, among others, the encoder configura-
tion may vary.
Also, for different models and manufacturers of DC motors, the P E and P P R data must be de-
termined according to what each manufacturer indicates. P P R = 64 for this case.
Finally, it must be considered that for the worked DC motor, the speed must be adjusted according
to the gear ratio given by the manufacturer, so the real speed in rpm is given by the following
expression:
v
speed =
GR
At this point, the speed curve can be plotted, or it can be adjusted so that the graphed data is
observed in percentage.
7. Labview Interface Considerations
The interface designed in Labview must allow graphing the motor’s output speed signal, ideally in
percentage but the physical magnitude in units of rpm can be worked with. Keep in mind:
- The encoder measures the motor’s rotation frequency before the gearbox, that is, the motor’s
rotation speed that is utilized is lower than the speed of the motor that receives the input signal
and that is measured by the encoder. The conversion or relationship is given by the gear ratio.
- The frequency delivered by the encoder is 16 times higher than the motor’s rotation frequency.
- The programming in Labview must take into account the scales of measured voltage, the motor’s
rotation frequency before the gearbox, and the real rotation speed in RPM.
8. Final Considerations
- A negative θ implies that the system has no delay.
7
Submission 2 Joint Project
Non-parametric identification
of the DC motor
- The manufacturer indicates that the motor operates inadequately for a voltage lower than 6 volts
in its supply, which is why the driver must be powered at 12 volts.
Identify the assigned motor. You must consult the datasheet of the motor in detail and determine
as many parameters as possible from the document.
9. Submission
Upload to moodle the support files for Labview, Matlab, and Simulink, and the upload of the
following results and discussion to the project webpage:
1. [1.5] The transfer functions found through the graphical method (POR and SOR).
For this, find 3 reaction curves corresponding to 3 steps of different magnitudes, and determine
corresponding parameters. Then average the parameters and obtain a resulting curve for each
method.
If one of the 3 curves yields values far removed or atypical compared to the other 2, this curve can
be discarded for the average, but the data must be reported.
If no curve resembles each other, meaning the parameters also do not resemble, choose a single
one and adjust the parameters to that curve.
2. [1.0] The validation, through a graphical comparison, of the dynamic behavior of the 3 obtained
models: POR model, SOR model, and the model obtained in the parametric identification. Keep
in mind that the graphical comparison must include real data, and the best model will be the one
that best fits the data.
To perform this procedure, keep in mind that real data and the step do not have initial conditions
equal to zero and that the input step is not unitary. There are two alternatives: the first is to adjust
the simulation so that the step signal and the model have the same conditions as the experiment,
or take the data and subtract an appropriate magnitude so that their initial conditions are zero.
3. [0.5] Obtain and write the POR and SOR models identified through the Matlab systemIdentifi-
cation tool, and their respective simulations.
4. [1.0] The validation, through a graphical comparison, of the dynamic behavior of the models
obtained through the Matlab systemIdentification tool. Compare the resulting model with the best
POR and SOR model selected according to the comparison in item 2.
8
Submission 2 Joint Project
Non-parametric identification
of the DC motor
5. [1.0] The discretization of the T.F.s selected in item 2 and, through a graphical comparison,
confirm that the selected model has the best behavior. Justify and report values of interest, such
as the sampling time and the method used.
The submission must be made through the Moodle platform according to the instructions and
established date. Consider:
In a ZIP or RAR, include file or files .m and simulink corresponding to the implementation
of the laboratory codes.
The files must be identified in the following way: • Matlab and simulink files: DC A1 TeamX.rar
10. Evaluation Rubrics
9
Submission 2 Joint Project
Non-parametric identification
of the DC motor
Niveles de dominio
Indicators
Excellent Good Acceptable Insufficient Deficient
Lab Presents re- Presents re- Presents Presents Presents inco-
Report sults of all sults of all incomplete incomplete rrect results
Content proposed proposed results of or incorrect for most acti-
(90 points) activities activities proposed ac- results of vities or does
clearly, with clearly, with tivities with proposed ac- not present
discussion of discussion of discussion of tivities with most acti-
results and results and results and discussion of vities, with
conclusions, conclusions conclusions results and discussion of
within the lacking depth lacking depth conclusions results and
stipulated or detail. or detail. lacking depth conclusions
deadline. Missing at or detail, or absent.
least one missing. Mis-
activity. sing 2 or more
activities.
95 45 30 15 0
Communication The report is The report’s The report’s The report’s The report’s
Skills (10 well-written, writing and writing and writing and writing and
points) with correct spelling re- spelling re- spelling re- spelling are
spelling. It quire minor quire general quire major deficient. It
uses an ar- corrections. It corrections. It corrections. It does not use
ticle format uses an article uses an article uses an article or uses an
with well- format, and format, and format, and article format
identified and the sections, the sections, the sections, inadequately,
numbered references, or references, references, with missing
sections, with numbering and numbe- and numbe- sections and
appropriate of tables, ring of tables, ring of tables, no proper
use of refe- figures, or figures, and figures, and order.
rences and formulas re- formulas re- formulas re-
numbering quire minor quire minor quire major
of tables, corrections. corrections. corrections.
figures, and
formulas.
10 8 6 4 0
10
