0% found this document useful (0 votes)
203 views8 pages

Process Control Final Exercises

Uploaded by

charlestw0514
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
203 views8 pages

Process Control Final Exercises

Uploaded by

charlestw0514
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Process Control Final Exercises (110/06/24)

1. An operator tests the dynamic behavior of a furnace in order to identify the transfer
function relating the furnace temperature (output) to the heat input. The operator
performs a serious of step increases/decreases in the amount of heat input and records
the temperature changes in the furnace. The system is initially at steady state, and the
test takes 10 h (see Fig. E5.30). Data are recorded for another 5 h after changes in the
input are no longer made.

(a) Does a visual inspection of the plots above provide enough information to
determine (i) the order of the system transfer function, (ii) the gain, and (iii) the time
constant(s)?
If yes, please provide the corresponding values and justify your answers. If an exact
determination is not possible for some of the above, justify why and discuss what a
good estimate would be. Please give the units where necessary.
(b) How should the test procedure be modified to improve the quality of the data and
make it easier (or possible) to identify the transfer function of the system from the
plot?
(c) What can you qualitatively infer about the following physical parameters of the
furnace (ignore the air inside): (i) mass, (ii) density, (iii) heat capacity, and (iv)
furnace height?
Do not perform any calculations; use words such as “large,” “small,” and
“significant.” Please justify your answer.

Ans
(a) 15 h 過後⇒ first order system
The gain (K)⇒ K=
0−(−1 ) [ K ]
=0.001 K /kW
0−(−1000 ) [kW ]
∵ 當 t=5時,系統達到 steady state。
∴ the time constant (τ) ⇒τ = (17−13 ) [h ] = 4[h] =0.8 h=2880 s
5 5
(b) 輸入的 step change 間隔應該拉大,大於 4 h 較好。
(c) ∵ 由 (a)得知 K 很小但 τ 很大⇒ m (mass)、 (density)、CP (heat capacity) 、H (furnace
height) 的值皆很大。

2. A step change from 15 to 31 psi in actual pressure results in the measured response
from a pressure-indicating element shown in Fig. E5.14.
(a) Assuming second-order dynamics, calculate all important parameters and write an
approximate transfer function in the form
'
R ( s) K
'
= 2 2
P ( s ) τ s +2 ζτs+1
where R' is the instrument output deviation (mm), P' is the actual pressure deviation
(psi).
(b) Write an equivalent differential equation model in terms of actual (not deviation)
variables.

Ans
(a) KM= y() – y(0)

⇒ K=
( 11.2−8 ) [mm]
=0.2 mm/ psi
( 31−15 ) [ psi]
a −π ζ / √(1−ζ ) 2

OS= =e
b

⇒ 12.7−11.2 =e
2
−π ζ / √(1−ζ )

11.2−8
⇒ ζ =0.234
2π τ
P= =2.3 [s ]
√ 1−ζ 2
⇒ τ =2.3 [ s ] × 1−0.234 =0.356[ s]∴ ' =

'
2
R (s ) 0.2
2
2π P (s ) 0.127 s + 0.167 s +1
'
R ( s) 0.2
(b) '
= 2
P ( s ) 0.127 s +0.167 s+1
inverse Laplace transform
0.127 R̈ ' + 0.167 R˙ ' + R '=0.2 P '
R̈ '=R̈ Ṙ '=Ṙ R '=R−8 P '=P−15 帶入上式
0.127 R̈+ 0.167 Ṙ+ R=0.2 P+5
→ R̈ +1.31 Ṙ+ 7.88 R=1.57 P+39.5

3. A process has the block diagram


−θs
2 2e
U (s ) ⟶ ⟶ ⟶ Y (s )
( 2 s +1 ) ( 0.4 s +1 ) ( 0.4 s +1 ) ( s +1 )
Derive an approximate first-order-plus-time-delay transfer function model.

Ans
−θs −θs −0.5 s −0.4 s −0.4 s −(θ+1.3 )s
4e 4e e e e 4e
G ( s )= 2
= =
( 2 s+ 1 )( s+1 ) ( 0.4 s+1 ) [ ( 2+0.5 ) s+1] 2.5 s+1

4. Consider the transfer function


−3 s
320 ( 1−4 s ) e
G ( s )=
24 s 2 +28 s +4
(a) What are the gain, time delay, time constants, poles and zeros of G(s)?
(b) Will the step response of this transfer function exhibit (i) inverse response or (ii)
oscillations?
Justify your answer.
Ans
−3 s −3 s
320 ( 1−4 s ) e 80 ( 1−4 s ) e
(a) G ( s )= =
24 s 2 +28 s +4 (6 s +1)(s +1)
∴K (gain)= 80;time delay= 3;τ1、τ2 (time constants)= 6、1
−1 1
poles= 、−1∧zeros=
6 4
(b) ∵如何收斂是看分子,而τ a=−4(¿ 0) ⇒ inverse response

5. A liquid-level control system can be configured in either of two ways: with a


control valve manipulating the flow of liquid into the holding tank (Fig. E8.5a), or
with a control valve manipulating the flow of liquid from the tank (Fig. E8.5b).
Assuming that the liquid-level transmitter always is direct-acting,
(a) For each configuration, what control action (direct or reverse) should a
proportional pneumatic controller have if the control valve is air-to-close?
(b) What if the control valve is air-to-open?
Ans
(a) 圖 a) if h↑,則 qi 需要↓。而對於 air-to-close 的控制閥,p (controller output)則需↑。∴ h↑
、p↑ ⇒ direct acting
圖 b) if h↑,則 qi 需要↑。而對於 air-to-close 的控制閥,p (controller output)則需↓。∴ h↑
、p↓ ⇒ reverse acting
(b) 圖 a) if h↑,則 qi 需要↓。而對於 air-to-open 的控制閥,p (controller output)則需↓。∴ h↑
、p↓ ⇒ reverse acting
圖 b) if h↑,則 qi 需要↑。而對於 air-to-open 的控制閥,p (controller output)則需↑。∴ h↑
、p↑ ⇒ direct acting

6. An electronic PI temperature controller has an output p of 12 mA when the set point


equals the nominal process temperature. The controller response to step change in the
temperature set point of 2.5 mA (equivalent to a change of 5°F) is shown below:

t (s) p (mA)
0- 8.0
0+ 6.7
20 6.0
60 4.7
80 4.0

Determine the controller gain Kc (mA/mA) and the integral time τI. Is the controller
reverse-acting or direct acting?
Ans
The instantaneous change at t = 0 is Kc and the slope of the response is Kc/τI.
consider a more general step change in the set point of magnitude M.

( )
P (s ) 1 Kc
=K c 1+ =K c +
E (s ) τI s τI s
M
E ( s )=
s
Kc M K c M
合並上兩式: P ( s )= + 2
s τI s
Kc M
p ( t ) =K c M + t for t≧0
τI
Kc M
Slope: ;instantaneous change at t = 0: K c M
τI
From the data given in the table, the initial instantaneous change is -1.3 mA and the
slope is -0.0335 mA/s for a step change of M = 2.5 mA.
−1.3 mA −1.3 mA
K c= = =−0.52
M 2.5 mA
Kc M mA
=−0.0335
τI s
Kc M −0.52 ( 2.5 mA )
τ I= = =38.8 s
mA mA
−0.0335 −0.0335
s s
Because K c is negative, we classify this controller as direct acting.

7. Consider proportional-only control of the stirred-tank heater control system in Fig.


E11.3. The temperature transmitter has a span of 50°F and a zero of 55°F. The
nominal design conditions are T =80 ℉ and T i=65 ℉ . The controller has a gain of 5,
while the gains for the control valve and current-to-pressure transducer are Kv = 1.2
(dimensionless) and KIP = 0.75 psi/mA, respectively. The time constant for the tank is
τ = 5 min. The control valve and transmitter dynamics are negligible. After the set
point is changed from 80 to 85°F, the tank temperature eventually reaches a new
steady-state value of 84.14°F, which was measured with a highly accurate
thermometer.
(a) What is the offset?
(b) What is the process gain K2?
(c) What is the pressure signal pt to the control valve at the final steady state?

Ans
(a) offset:5-4.14=0.86℉
(b) K m = 20−4 =0.32 mA
50 ℉
Kc=5
KIP=0.75
KV=1.2
τ =5
' 5
T sp ( s )=
S
K2
'
T (s) K m K c K IP K V
τs+1
'
=
T sp ( s ) K2
1+ K m K C K IP K V
τs+ 1
K2
0.32× 5 ×0.75 ×1.2 ×
' 5 s +1 5
T ( s )= ×
K2 s
1+0.32 ×5 × 0.75× 1.2×
5 s+1
K2
1.44 ×
5 s+1 5
¿ ×
K2 s
1+ 1.44 ×
5 s+ 1
1.44 × K 2 5
¿ ×
5 s+1+1.44 × K 2 s
7.2 K
lim ( s ×T )= 1+1.44 2K
'
(s ) =4.14
s→0 2

4.14 + 4.14 ( 1.44 K 2 )=7.2 K 2


4.14 +5.96 K 2=7.2 K 2
4.14 ℉
K 2= =3.34
7.2−5.96 psi
(c)
'
P t( s ) K IP K c K m
'
=
T sp ( s ) 4.81
1+
5 s +1
0.75 ×5 ×0.32
¿
4.81
1+
5 s+1
1.2
¿
4.81
1+
5 s +1
' 6 s +1 5
P t ( s) = ×
5 s +1+4.81 s
30 s +6
P t( t )=lim ( s × P t( s) )=
' '

s →0 5 s+5.81
6
=1.032psi
'
P t ( ∞) =
5.81

8. A block diagram of a closed-loop system is shown in Fig. E11.10.

(a) Derive a closed-loop transfer function for disturbance changes, Y(s)/D(s).


(b) For the following transfer functions, what values of Kc will result in a stable
closed-loop system?
4 1
G 1 ( s )=5 G 2 ( s )= K m=1 G 3 ( s ) =
2 s+1 s−1
Ans
(a)Derive CLTF:
Y=Y3+Y2=G3Z+G2P
Y=G3(D+Y1)+G2KCE
Y=G3D+D3D1KCE+G2KCE
Y=G3D+(G3G1KC+G2KC)E E=-KmY
Y=G3D-KC(G3G1+G2)KmY
[1+KC(G3G1+G2)Km]Y=G3D
Y G3
=
D 1+ K C ( G3 G1+G 2) K m

(b) Characteristic Equation:


1+KC(G3G1+G2)Km=0

1+ K C
[ 5
+
4
s−1 2 s +1
=0
]
1+ K C
[ ( s−1 ) ( 2 s+1 ) ]
5 ( 2 s+1 ) + 4 ( s−1 )
=0

( s−1 )( 2 s +1 ) + K C [ 5 (2 s+1 ) + 4 ( s−1 ) ] =0


2s2-s-1+Kc (10s+5+4s-4)=0
2s2+(14Kc-1)S+(Kc-1)=0
必要條件: Kc > 1/14 and Kc > 1
⇒Kc > 1 result in a stable closed-loop system.

Common questions

Powered by AI

For the transfer functions G1(s) = 5, G2(s) = 4/(2s+1), G3(s)= 1/(s-1), and considering the characteristic equation, the closed-loop system will remain stable for Kc values greater than 1. This is calculated by solving the characteristic polynomial derived from the feedback loop, ensuring all pole roots of the system's equation have negative real parts for stability .

A block diagram with time delay and multiple poles, such as G(s)=4 e^(-θs)(2s+1)(s+1)(0.4s+1)^2, typically indicates a system with considerable lag and complexity in its response. Time delays can significantly destabilize closed-loop systems or introduce non-minimum phase behavior, and multiple poles usually suggest slower dynamics and potential stability issues. Analyzing this diagram helps in predicting potential oscillations or sluggish response times based on the interplay of delays and pole locations .

The damping ratio (ζ) can be predicted using the overshoot formula: OS = (peak value - new steady state) / (new steady state - initial value). For a step change from 15 to 31 psi, calculate OS and solve for ζ using OS = e^(-πζ/√(1-ζ²)). The natural frequency (ω_n) can be calculated using the period of oscillations: τ = (2π/ω_n)√(1-ζ²). Using these relationships, the damping ratio and natural frequency are derived theoretically from the given system response data .

If a liquid-level control system with an air-to-close valve is configured to manipulate the inflow, a direct acting controller is required because an increase in level would need an increase in controller output to reduce the inflow. In contrast, if the valve is configured to control outflow, a reverse acting controller is needed. For an air-to-open valve, the actions are reversed: a rise in level would need an opposite action compared to the air-to-close setups .

A transfer function might exhibit an inverse response if its numerator has a zero with a negative real part, which essentially causes an initial response opposite to the long-term steady state after a step input. Characteristics include early decrease before rising in response to a positive input change, which is confirmed by deriving the transfer function: such as zeros located at the rightmost part of the complex plane numerically .

To improve data quality for identifying a system's transfer function, the interval between step changes in the input should be larger, preferably greater than the identified time constant. This allows the system to fully respond to each change, minimizing the overlap in responses, thus making the dynamic behavior clearer and easier to analyze .

The necessary parameters to determine the transfer function of the furnace system include the order of the system, the gain, and the time constant(s). Visual inspection of plots may not always provide enough information to accurately determine these parameters because visual interpretation lacks precision, especially in complex systems where multiple factors can influence the output. The system's behavior might be influenced by noise, nonlinearity, or other disturbances not readily apparent in visual data alone .

Given the small gain and large time constant inferred from the system's dynamics, it can be qualitatively inferred that the mass (m), density (ρ), heat capacity (C_p), and furnace height (H) are all likely large. These large physical parameters contribute to a system that exhibits significant inertia and takes a long time to respond to input changes .

Modifications in a PI controller's integral time (τI) and gain (Kc) can enhance stability and response by carefully tuning τI to reduce steady-state error without introducing instability and adjusting Kc to ensure sufficient responsiveness without causing overshoot. For the given system, calculated values like τI = 38.8 s indicate a balanced approach between transient and steady-state performance, requiring adjustments based on step response observations indicating under-damping or excessive response times .

Control valve and transmitter dynamics significantly impact offset and stability; however, when considered negligible, as in a stirred-tank system, it results in a direct influence by the controller gain setting, leading to a steady-state offset due to proportional-only control. The precise calculations of the offset illustrate a reduced ability to perfectly match the setpoint if proportional control lacks integral action balancing gains and system responsiveness adequately .

You might also like