Question 5.8.3: Set up the equations of motion for the actuator shown in Fig...
Set up the equations of motion for the actuator shown in Fig. 5.15. Assume that all magnetic materials are linear with μ >> μ_{0}. The nonmagnetic sleeves of thickness s provide low friction support for the moving plunger [4].

Learn more on how we answer questions.
Set up a coordinate system with f_{fld} positive in the direction of increasing x as shown. As the motion is linear, we use the first-order system
of equation (5.86)
\frac{di(t)}{dt} = \frac{1}{L(x)} [V_{s}(t) – i(t) (R+R_{coil}) – \frac{∂Λ(i,x)}{∂x} v(t)]
\frac{dv(t)}{dt} =\frac{1}{m} F(i,x)
\frac{dx(t)}{dt} = v(t). (5.173)
To solve Eq. (5.173) we need expressions for L(x), ∂Λ(i, x)/∂x and F(i, x). As the independent variables are i and x, we work with the coenergy W_{fld}^{c}(i, x). From Eq. (5.139) we have
Eq. (5.139): W_{fld}^{c} = \left\{\begin{matrix} \frac{∂Λ(i,x)i}{2}=\frac{L(x)i^{2}}{2} \text{(linear motion)} \\ \frac{∂Λ(i,θ)i}{2}=\frac{L(θ)i^{2}}{2} \text{(rotational motion)}\end{matrix} \right.
W_{fld}^{c} ( i,x) = \frac{Λ(i,x)i}{2}.
= \frac{L(x)i^{2}}{2}. (5.174)
Inductance: We determine the inductance using Eq. (3.75),
Eq. (3.75): L = \left\{\begin{matrix} \frac{1}{I^{2}} \int_{V} A⋅J dv \\ \frac{1}{I^{2}} \int_{V} B⋅H dv. \end{matrix} \right. (linear system)
L = \frac{1}{i^{2}} \int_{V} B ⋅ H dv. (5.175)
As the core and plunger have a high permeability, H is negligible in these elements and Eq. (5.175) reduces to an integration over the gap region and the nonmagnetic sleeve regions,
L = \frac{1}{i^{2}} \int_{V_{g}} B ⋅ H dv + \frac{2}{i^{2}}\int_{V_{s}} B⋅H dv, (5.176)
where V_{g} and V_{s} are the volumes of the gap and sleeve regions, respectively. The factor of 2 in the second integral takes into account the identical integrations over each sleeve region. Notice that V_{g} = A_{g}x and V_{s} = A_{s}s, where A_{g} and A_{s} are the cross-sectional areas of the gap and sleeve, respectively. We need to determine H_{g} and H_{s}. To this end, apply Eq. (3.141) to the dotted path in Fig. 5.15. This gives
Eq. (3.141): \oint_{C} H⋅ dl = \sum\limits_{i=1}^{m} H_{i} l_{i} = I_{tot}.
H_{g} x + H_{s} s = ni. (5.177)
Further, as there is no flux leakage Φ_{g} = 2Φ_{s},or
B_{g}A_{g} = 2B_{s} A_{s} . (5.178)
Combining Eqs. (5.177) and (5.178) with B_{g} = μ_{0}H_{g} and B_{s} = μ_{0}H_{s} gives
H_{g} = \frac{ni}{(x+(A_{g}/2A_{s})s)}, (5.179)
and
H_{s} = \frac{A_{g}}{2A_{s}} \frac{ni}{(x+(A_{g}/2A_{s})s)}. (5.180)
Next, substitute Eqs. (5.179) and (5.180) into Eq. (5.176), which gives the inductance
L(x) = \frac{μ_{0}n^{2}A_{g}}{(x+(A_{g}/2A_{s})s)} (5.181)
Coenergy: Once we know the inductance, the coenergy is obtained easily from Eq. (5.159)
Eq. (5.159): W_{fld}^{c} ( i,x) = \frac{L(x)i^{2}}{2}.
W_{fld}^{c} (i, x) = \frac{μ_{0}i^{2}n^{2}A_{g}}{2[x+(A_{g}/2A_{s})s]}. (5.182)
Use Eqs. (5.137) and (5.138) to obtain Λ and f_{fld}. We find that
Eq.(5.138):
f_{fld} = \frac{∂W_{fld}^{c} (i, x)}{∂i} (linear motion)
T_{fld} = \frac{∂W_{fld}^{c} (i, θ)}{∂θ} (rotational motion)
Λ = \frac{∂W_{fld}^{c} (i, x)}{∂i}
= \frac{μ_{0}in^{2}A_{g}}{[x+(A_{g}/2A_{s})s]}. (5.183)
and
f_{fld} = \frac{∂W_{fld}^{c} (i, x)}{∂x}
= \frac{μ_{0}i^{2}n^{2}A_{g}}{2[x+(A_{g}/2A_{s})s]^{2}}. (5.184)
Notice that f_{fld} is negative, which implies that the direction of f_{fld} is opposite to the direction of increasing x (i.e., opposite to the direction of increasing air gap). The total force F(i, x) is a sum of the forces due to the field and the spring,
F(i, x) = f_{fld}(i,x) + f_{s}(x), (5.185)
where
f_{s}(x) = -k(x -x_{0}). (5.186)
Here, x_{0} is the initial position of the plunger, which is assumed to be the equilibrium position for the spring. When x > x_{0} (x < x_{0}), f_{s}(x) tends to move the mass into (out of) the circuit. Finally, substitute Eqs. (5.181), (5.184), (5.185), and (5.186) into Eq. (5.173) to obtain
\frac{di(t)}{dt} = \frac{(x+(A_{g}/2A_{s})s)}{μ_{0} n^{2}A_{g}} \{ V_{s}(t) – i(t) (R+R_{coil}) + \frac{μ_{0}i n^{2}A_{g}}{[x+(A_{g}/2A_{s})s]^{2}} v(t) \}
\frac{dv(t)}{dt} = -\frac{1}{m} \{ \frac{μ_{0}πi^{2} n^{2}d}{2g} \frac{h^{2}}{(2h-x)^{2}}+ k(x-x_{0}) \}
\frac{dx(t)}{dt} = v(t). (5.187)
This nonlinear first-order system is solved subject to the initial conditions x(0) = x_{0}, v(0) = v_{0} and i(0) = i_{0}.