2. Materials and Methods
In dealing with the replaceable interface mast, the reconfigurable modules within the mast are considered rigid bodies. As shown in
Figure 1, two fixed points, p and q, and two unit vectors, u and v, are used to establish a natural coordinate system, x-y-z, on the rigid module.
Let Vr represent the volume of the reconfigurable rigid body module and ρ represents its density. The expression for the kinetic energy of the reconfigurable module can be calculated by Eq.1.
Where Vr is the Volume of reconfigurable modules, M is the quality matrix of the reconfigurable module, ; A is the shape functions of the rigid body module, is a generalized coordinate system, is a position vector in the global coordinate system.
The virtual work of inertial forces for reconfigurable modules can be calculated by Eq.2.
The external virtual work of the reconfigurable module can be written as Eq.3.
In Eq.3, QF and QM respectively represent the generalized force columns of external force and external moment in the natural coordinate system. The virtual work done by the force F acting at any point N on the rigid body module can be expressed as Eq.4.
Where AN is the shape function corresponding to point N.
The virtual work of internal forces in the rigid body module is zero, hence, we can infer from the principle of virtual work that:
Where is the generalized force due to external forces, is the generalized force corresponding to the moment M.
Establish the intrinsic constraint equations for the reconfigurable module in the natural coordinate system as Eq.6.
The dynamic equations for the reconfigurable rigid body module in the absence of external constraints can be obtained in Eq.7.
Where .
The interface on the mast of the replaceable interface can be equivalently modeled as a cylindrical shell. The shell used in this paper has a thickness of h=2mm and a mid-surface radius of R=41mm, thus the interface can be represented by thin shell elements. The schematic diagram of the shell element deformation is shown in
Figure 2.
The present study employs the gradient-reduced ANCF thin shell element [
19] to discretize the finite element of the flexible interface's equivalent cylindrical shell model. The deformation along the thickness of the thin shell is neglected, and the overall deformation of the shell element is described through the stretching and shearing deformations on the mid-surface of the cylindrical shell element, as well as the bending deformation of the element.
2.1. Description of the movement of the shell unit
The nodal coordinates of the shell element, denoted as "e," are composed of the position coordinates and gradient coordinates of four points A, B, C, and D on the neutral surface. The nodal coordinates of each point can be expressed as Eq.8.
The first term in the expression represents the position coordinate array, while the second to the third terms represent the gradient coordinate array. Here, i represents the points A, B, C, and D, and the specific expressions for each term are as Eq.9.
Therefore, the expression for the nodal coordinates of the element is as Eq.10.
The position vector of any point Q on the neutral surface of the shell element in the global coordinate system is given by Eq.11.
In the expression, S represents the shape function of the shell element. The expression for S is Eq.12.
In the Eq.12. I3 denotes the third-order identity matrix.
Therefore, the expressions for the velocity and acceleration of any point on the shell element can be obtained from the equation in Eq.13.
2.2. Deformation modeling of shell units
As shown in
Figure 2 and
Figure 3, the schematic diagram represents a surface within the shell element, with the global coordinate system denoted as X-Y-Z and the α-β-γ system representing the curvilinear coordinates on the surface of the shell element.
The g10-g20-n0 and e10-e20-n0 correspond to the local curvilinear coordinate system and the Cartesian coordinate system on the shell element, respectively.
The vector of microsegments between two adjacent points Q and P on the midplane before and after the deformation can be written as Eq.14.
The definition of the basis vectors for two sets of local coordinate systems on the mid-surface is as Eq.15.
Coordinate transformation yields can be expressed as Eq.16.
Here, Tθ represents the transformation matrix from the Cartesian coordinate system to the curvilinear coordinate system, where θ denotes the angle between the coordinate axes of the curvilinear coordinate system.
The formula for calculating the Lagrange strain tensor can be derived from continuum mechanics as Eq.17.
The expression for the mid-surface Lagrange strain tensor is obtained as Eq.18.
Since
εmid has symmetry,
εmid can be rewritten as Eq.19.
From
Figure 2-1, the position vector of any point Qʹ on the external surface of the shell unit can be expressed as Eq.20.
Then the basis vector of the surface coordinate system at point Qʹ can be calculated by Eq.21.
Similarly, the length of the micro-segment arc between two adjacent points Qʹ and Pʹ on the external surface before and after the deformation can be obtained by Eq.22.
The Lagrange strain tensor for the shell element can be calculated by Eq.23.
Where is unit bending strain.
By neglecting higher-order terms in z, we obtain Eq.24. and Eq.25.
2.3. Flexible interface dynamics model
Let the volume of the shell unit be Ve, and then the virtual work of the inertial force on the shell unit can be calculated by Eq.26.
Where is the mass matrix of the shell unit.
Therefore, the virtual work of inertia force for the flexible interface can be calculated by Eq.27.
Where n is the number of shell units after interface discretization; Ve is the volume of the i-th shell unit.
is The generalized coordinates of the interface that can be calculated by Eq.28.
where B
j is denoted as the boolean matrix corresponding to each shell cell.
As the deformation of the shell element is decomposed into the bending deformation of the shell element and the deformation on the mid-surface, the strain energy of the shell element can be decomposed into the bending strain energy (Ub) and the mid-surface deformation strain energy (Umid), with the following calculation formulas.
Where
E is the third-order modulus of the elasticity matrix; K is the equivalent transformation matrix of the bending strain tensor.
Where is Poisson's ratio.
The derivative of the strain energy for the generalized coordinates yields the elastic forces on the neutral surface.
The bending strain energy is derived for the generalized coordinates to obtain the bending elastic force as Eq.33.
Then the shell unit elastic force is
, so the internal force virtual work of the shell unit is
, and the internal force virtual work of the flexible interface can be calculated by Eq.34 and Eq.35.
Let the generalized force acting on the interface be
Qa, then its corresponding external virtual work can be calculated by Eq.36.
Then, from the principle of virtual work, we can obtain the equation as Eq.37.
Where .
Then the kinetic equation of the flexible interface without external constraints can be expressed as Eq.36.
2.4. Dynamics model of the mast deployment process with replaceable interface
The schematic diagram of the interchangeable interface mast system, known as the "three interfaces and four modules," is illustrated in
Figure 3.
When deriving the dynamics equations for the expansion process of the interchangeable interface mast system, the following assumptions are made:
(1) The reconfigurable module is considered as a fixed constraint between the reconfigurable module and the flexible interface;
(2) The rotation constraint is between the reconfigurable module and the body of the rover (base);
(3) The mast deployment process does not consider the effect of gravity;
(4) The body of the rover is considered a rigid body, and the body of the rover is at rest when the mast is deployed.
After establishing the dynamic equations for each part separately, it is necessary to establish the dynamic equations of the mast system by simultaneously combining the dynamic models of each part through the system's constraint equations. The constraints of the mast system include the articulated constraints between the reconfigurable modules and the inspector body, the fixed constraints between the reconfigurable modules and the flexible interfaces, as well as the intrinsic constraints introduced when modeling the reconfigurable modules using NCF.
The reconfigurable modules, which are described by the NCF method to depict the motion of rigid bodies, possess 12 generalized coordinates. Consequently, it is necessary to impose 6 constraints, including the mutual orthogonality constraints between three basis vectors and the length constraints of the basis vectors. The expression for the constraint equations of the rigid body is as Eq.39.
Where d is the side length of the reconstructed module (m).
2.5. Fixed Constraints
The fixed constraint limits the position between the origin of the coordinate system on the shell unit and the nodes on the rigid module and the rotation of the interchangeable interface for the rigid module, as shown in
Figure 4.
The gradient vector along the x-direction at the nodes A and B during the unfolding process is always perpendicular to the vector u. As Eq.40.
Where is the radius of the neutral surface of the interface.
2.6. Rotation constraints
The rotational constraints between the reconfigurable module and the base are illustrated in
Figure 4-2. Here, the x-y-z axes represent the local coordinate system of the rigid body, while the X-Y-Z axes represent the global coordinate system. Thus, the center of mass of the rigid body module coincides with the origin of the global coordinate system. Additionally, during the rotation of the rigid body module, the ox and oy axes are always perpendicular to the OZ axis. Hence, the constraint equations can be expressed as Eq.41.
2.7. Mast system dynamics equation
For a system of replaceable interface masts with m flexible interfaces and n reconfigurable modules, the dynamical equations in the unfolding process can be expressed as Eq.41.
Let
be the generalized coordinates of the total interchangeable interface mast system, then the above equation can be rewritten as Eq.43.
,.
The replaceable interface mast system has a total of 1 rotational constraint and 2m fixed constraints, and the constraint equation of the mast system can be obtained by Eq.44.
Where is the total generalized coordinate of the interface contained in the mast system.
The dynamical equations in matrix form during the unfolding of the mast system can be obtained by associating the Lagrange multiplier method. As Eq.45.
Where the Jacobi matrix of the system constraint equations for the generalized coordinates of the system; λ is the Lagrange multiplier array.
The expression obtained as Eq.46. by taking the second-order derivative of the constraint equation
for time and organizing.