profile pic
⌘ '
raccourcis clavier

See lab notes

prelab.

The transfer function of a DC electric motor with respect to angular velocity is:

Gω(s)=Ω(s)V(s)=Aτs+1G_{\omega}(s) = \frac{\Omega(s)}{V(s)} = \frac{A}{\tau s + 1}

Where

  • AA and τ\tau are positive, real-valued constants
  • V(s)V(s) and Ω(s)\Omega(s) are voltage and angular velocity as function of ss in the Laplace domain.

Note that Ω(s)L{ω(t)}\Omega(s) \coloneqq \mathcal{L}\{\omega(t)\}

Q1.

We will now develop a formula for the motor DC gain constant AA in terms of a step input change ΔV\Delta V and step output change Δω\Delta \omega

problem a.

Using the final value theorem, find an expression for the steady state value of ω(t)\omega(t) when a step input of amplitude VxV_x is applied.

note: τ>0\tau > 0 so the pole of Gω(S)G_{\omega}(S) at s=1τs=-\frac{1}{\tau} is in the open Left Half Plane (LHP) so the system is stable.

Given that the final value theorem, the steady state value of f(t)f(t) is given by:

limtf(t)=lims0sF(s)\lim_{t \to \infty} f(t) = \lim_{s \to 0} s F(s)

When a step input V(s)=VxsV(s) = \frac{V_x}{s} is applied, the output in the Laplace domain is given by:

Ω(s)=Gω(s)V(s)=Aτs+1Vxs limtω(t)=lims0sΩ(s)=lims(AVxτs+1)\begin{align*} \Omega(s) = G_{\omega}(s) \cdot V(s) &= \frac{A}{\tau s + 1} \cdot \frac{V_x}{s} \\\ \lim_{t \to \infty} \omega(t) = \lim_{s \to 0} s \Omega(s) &= \lim_{s \to \infty} (\frac{A \cdot V_x}{\tau s + 1}) \end{align*}

Since τ>0\tau > 0 when s0s \to 0:

limtinfω(t)=AVxτ0+1=AVx\lim_{t \to \inf} \omega(t) = \frac{A \cdot V_x}{\tau \cdot 0 + 1} = A \cdot V_x

The steady-state value of ω(t)\omega(t) under a step amplitude of VxV_x is directly proportional to the product of DC gain and step input VxV_x

problem b.

Give the expression for ω(t)\omega (t) in response to a step input VxV_x at time t=0t=0. Assume a non-zero initial condition for ω(t)\omega (t), i.e., ω(t)=ω0\omega(t) = \omega_0.

note: the response due to a non-zero initial condition ω0\omega_0 can be modeled as the response due to input v(t)=ω0δ(t)v(t) = \omega_0 \delta(t) where δ(t)\delta(t) is the impulse function. Since Gω(s)G_{\omega}(s) is a linear system, the response to step input VxV_x with non-zero initial condition ω0\omega_0 is just the sum of the responses due to the step and the initial condition.

The zero-state response to step input VxV_x is given by the inverse Laplace transform of Gω(s)V(s)G_{\omega}(s) \cdot V(s):

ωzs(t)=AVx(1etτ)\omega_{zs}(t) = AV_x \cdot (1 - e^{-\frac{t}{\tau}})

The zero-input response to initial condition ω0\omega_0 can be modeled as the response to an impulse input ω0δ(t)\omega_0 \delta(t). The Laplace transform of the impulse response is:

Ωzi(s)=Aω0τs+1\Omega_{zi}(s) = \frac{A \cdot \omega_0}{\tau s + 1}

Which the zero-input response is given by: ωzi(t)=ω0etτ\omega_{zi}(t) = \omega_0 \cdot e^{-\frac{t}{\tau}}

The total response ω(t)\omega(t) is the sum of the zero-state and zero-input responses (due to linearity):

ω(t)=ωzs(t)+ωzi(t)=AVx(1etτ)+ω0etτ\omega(t) = \omega_{zs}(t) + \omega_{zi}(t) = AV_x \cdot (1 - e^{-\frac{t}{\tau}}) + \omega_0 \cdot e^{-\frac{t}{\tau}}

problem c.

For ω(t)\omega(t) in computed in part b, what is the limtω(t)\lim_{t \to \infty} \omega(t)? How does it compare to result in part a?

limtω(t)=limt(AVx(1etτ)+ω0etτ)\lim_{t \to \infty} \omega(t) = \lim_{t \to \infty} (AV_x \cdot (1 - e^{-\frac{t}{\tau}}) + \omega_0 \cdot e^{-\frac{t}{\tau}})

Since etτ0e^{-\frac{t}{\tau}} \to 0 as tt \to \infty, the steady-state value of limtω(t)\lim_{t \to \infty} \omega(t) is:

limtω(t)=AVx(10)+0=AVx\lim_{t \to \infty} \omega(t) = AV_x (1-0) + 0 = AV_x

We see that the steady-state value of ω(t)\omega(t) is the same. The initial condition ω0\omega_0 does not affect the steady-state value of ω(t)\omega(t), only influence transient response.

problem d.

Now assume that you run the motor with an initial step input of VminV_{\text{min}} until time t0t_0. At time t0t_0, assume that the system has reached steady state and the step input is changed to VmaxV_{\text{max}} at time t0t_0. In other words, the system input will take the form

v(t)={vminif 0t<t0 vmaxif tt0v(t) = \begin{cases} v_{\text{min}} & \text{if } 0 \leq t < t_0 \\\ v_{\text{max}} & \text{if } t \geq t_0 \end{cases}

where t0τt_0 \gg \tau and VmaxV_{\text{max}} and V_min V\_{\text{min}} may be non-zero.

Use the results above to show that:

A=ΔωΔVA = \frac{\Delta \omega}{\Delta V}

where ΔV=VmaxVmin\Delta V = V_{\text{max}} - V_{\text{min}} and Δω=ωssω0\Delta \omega = \omega_{ss} - \omega_0 where ω0\omega_0 is the steady-state response to a constant input VminV_\text{min} and ωss\omega_{ss} is the steady-state response to the input VmaxV_\text{max}

For input VminV_{\text{min}}, the steady-state response is ω0=AVmin\omega_0 = A \cdot V_{\text{min}}

Similarly, for input VmaxV_{\text{max}}, the steady-state response is ωss=AVmax\omega_{ss} = A \cdot V_{\text{max}}

Thus, the change in steady-state response is

Δω=ωssω=AVmaxAVmin=AΔV\Delta \omega = \omega_{ss} - \omega = A \cdot V_{\text{max}} - A \cdot V_{\text{min}} = A \cdot \Delta V

Thus A=ΔωΔVA = \frac{\Delta \omega}{\Delta V}

Q2.

Using the formula derived in Q1, use the following graphs to calculate A for this system.

Given A=ΔωΔVA = \frac{\Delta \omega}{\Delta V}, from the graph Vmin=1V_{\text{min}} = 1 and Vmax=5V_{\text{max}} = 5 and ω0=5\omega_0 = 5 and ωss=25\omega_{\text{ss}} = 25, A=5A = 5

Q3.

For a first order system, the time it takes a step response to reach 63.2% of its steady state value (t1t0t_1 − t_0 in Fig. 1) is the response’s time constant τ\tau . i.e., at time t1,ω(t1)=0.632Δω+ω0t_1, \omega(t_1) = 0.632\Delta \omega + \omega_0. Find the time constant τ\tau for the above system.

ω(t1)=0.632Δω+ω0=0.632(255)+5=17.64\omega(t_1) = 0.632\Delta \omega + \omega_0 = 0.632 \cdot (25 - 5) + 5 = 17.64

From the graph, τ0.8sec\tau \approx 0.8 \sec

Q4.

Using AA and τ\tau calculated in Q2 and Q3, find the transfer function in terms of ss

Gω(s)=Ω(s)V(s)=Aτs+1=30.05s+1G_{\omega}(s) = \frac{\Omega(s)}{V(s)} = \frac{A}{\tau s + 1} = \frac{3}{0.05s + 1}

Q5.

The system quickly rises to a steady-state value without any oscillation, which suggests a first-order system. The transfer function:

G(s)=Kτs+1G(s) = \frac{K}{\tau s + 1}

Q6.

Deriving a transfer function experimentally and then use simulation software to design is preferable in situations where experimenting directly with the plant poses high risks, incurs excessive costs, requires downtown. For example, a chemical processing plant, probably we don’t want to experiment with the actual system, since it could lead to dangerous chemical reactions, waste of materials. Using simulation engineers can safely and cost-effectively design and test control strategies before implementing in real system.

Conversely, deriving transfer function experimentally and then using simulation software to design is not preferable in situations where the plant is simple and safe to experiment with, and the cost of experimenting is low. For instance, a small educational laboratory setup with low-cost components and minimal hazardous concerns would mean it might be more practical and education to design and calibrate the controller directly through experimentation and observe behaviours in real-time.


lab.

  • Check to change the TCP address of the model URI from QUARC > Settings > Preferences > Model
G(w)=1.8674680.027947s+1G(w) = \frac{1.867468}{0.027947 * s + 1}